Web Data Extraction, Applications and Techniques: A Survey

EMILIO FERRARA Indiana University Bloomington, US and PASQUALE DE MEO and GIACOMO FIUMARA University of Messina, Italy and ROBERT BAUMGARTNER Lixto Software GmbH, Austria

arXiv:1207.0246v1 [cs.IR] 1 Jul 2012

Web Data Extraction is an important problem that has been studied by means of dierent scientic tools and in a broad range of application domains. Many approaches to extracting data from the Web have been designed to solve specic problems and operate in ad-hoc application domains. Other approaches, instead, heavily reuse techniques and algorithms developed in the eld of Information Extraction. This survey aims at providing a structured and comprehensive overview of the research eorts made in the eld of Web Data Extraction. The l rouge of our work is to provide a classication of existing approaches in terms of the applications for which they have been employed. This dierentiates our work from other surveys devoted to classify existing approaches on the basis of the algorithms, techniques and tools they use. We classied Web Data Extraction approaches into categories and, for each category, we illustrated the basic techniques along with their main variants. We grouped existing applications in two main areas: applications at the Enterprise level and at the Social Web level. Such a classication relies on a twofold reason: on one hand, Web Data Extraction techniques emerged as a key tool to perform data analysis in Business and Competitive Intelligence systems as well as for business process re-engineering. On the other hand, Web Data Extraction techniques allow for gathering a large amount of structured data continuously generated and disseminated by Web 2.0, Social Media and Online Social Network users and this oers unprecedented opportunities of analyzing human behaviors on a large scale. We discussed also about the potential of cross-fertilization, i.e., on the possibility of re-using Web Data Extraction techniques originally designed to work in a given domain, in other domains. Categories and Subject Descriptors: H.2.8 [Database Management]: Database Applications Data mining; H.3.3 [Information Storage and Retrieval]: Information Search and Retrieval Retrieval models; I.2.5 [Articial Intelligence]: Programming Languages and SoftwareExpert system tools and techniques General Terms: Algorithms, Design, Theory Additional Key Words and Phrases: Web Information Extraction, Web Data Mining

E-mail addresses: [email protected] (E. Ferrara); [email protected] (P. De Meo); [email protected] (G. Fiumara); [email protected] (R. Baumgartner) Permission to make digital/hard copy of all or part of this material without fee for personal or classroom use provided that the copies are not made or distributed for prot or commercial advantage, the ACM copyright/server notice, the title of the publication, and its date appear, and notice is given that copying is by permission of the ACM, Inc. To copy otherwise, to republish, to post on servers, or to redistribute to lists requires prior specic permission and/or a fee. c 2012 ACM 0360-0300/12/0700-0001 $5.00
ACM Computing Surveys, Vol. V, No. N, July 2012, Pages 154.

Emilio Ferrara et al.

Contents 1 Introduction 1.1 Challenges of Web Data Extraction 1.2 Related works . . . . . . . . . . . . 1.3 Our contribution . . . . . . . . . . 1.4 Organization of the survey . . . . . 3 3 4 5 6 6 7 7 8 12 13 16 18 20

techniques . . . . . . . . . . . . . . . . . . . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

2 Techniques 2.1 Tree-based techniques . . . . . . . . . . . . . . . . . . . . . 2.1.1 Addressing elements in the document tree: XPath . 2.1.2 Tree edit distance matching algorithms . . . . . . . 2.2 Web wrappers . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.1 Wrapper generation and execution . . . . . . . . . . 2.2.2 The problem of wrapper maintenance . . . . . . . . 2.3 Machine Learning approaches . . . . . . . . . . . . . . . . . 2.3.1 Hybrid systems: learning-based wrapper generation

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

3 Web Data Extraction Systems 21 3.1 Layer cake comparisons . . . . . . . . . . . . . . . . . . . . . . . . . 23 4 Applications 4.1 Enterprise Applications . . . . . . . . . . . . . . . . . . . . . . . . 4.1.1 Context-aware advertising . . . . . . . . . . . . . . . . . . . 4.1.2 Customer care . . . . . . . . . . . . . . . . . . . . . . . . . 4.1.3 Database building . . . . . . . . . . . . . . . . . . . . . . . 4.1.4 Software Engineering . . . . . . . . . . . . . . . . . . . . . . 4.1.5 Business Intelligence and Competitive Intelligence . . . . . 4.1.6 Web process integration and channel management . . . . . 4.1.7 Functional Web application testing . . . . . . . . . . . . . . 4.1.8 Comparison shopping . . . . . . . . . . . . . . . . . . . . . 4.1.9 Mashup scenarios . . . . . . . . . . . . . . . . . . . . . . . . 4.1.10 Opinion mining . . . . . . . . . . . . . . . . . . . . . . . . . 4.1.11 Citation databases . . . . . . . . . . . . . . . . . . . . . . . 4.1.12 Web accessibility . . . . . . . . . . . . . . . . . . . . . . . . 4.1.13 Main content extraction . . . . . . . . . . . . . . . . . . . . 4.1.14 Web (experience) archiving . . . . . . . . . . . . . . . . . . 4.1.15 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 Social Web Applications . . . . . . . . . . . . . . . . . . . . . . . . 4.2.1 Extracting data from a single Online Social Web platform . 4.2.2 Extracting data from multiple Online Social Web platforms 4.3 Opportunities for cross-fertilization . . . . . . . . . . . . . . . . . . 28 30 30 31 31 32 32 32 33 33 34 34 35 35 35 36 36 36 38 41 44

. . . . . . . . . . . . . . . . . . . .

5 Conclusions and Ongoing Work 46 5.1 A glance on the future . . . . . . . . . . . . . . . . . . . . . . . . . . 46

ACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

1.

INTRODUCTION

Web Data Extraction systems are a broad class of software applications targeting at extracting information from Web sources like Web pages [Laender et al. 2002; Baumgartner et al. 2009]. A Web Data Extraction system usually interacts with a Web source and extracts data stored in it: for instance, if the source is a HTML Web page, the extracted information could consist of elements in the page as well as the full-text of the page itself. Eventually, extracted data might be post-processed, converted in the most convenient structured format and stored for further usage [Zhao 2007; Irmak and Suel 2006]. Web Data Extraction systems nd extensive use in a wide range of applications like the analysis of text documents at disposal of a rm (like e-mails, support forum, technical and legal documentation, and so on), Business and Competitive Intelligence [Baumgartner et al. 2005], crawling of Social Web platforms [Catanese et al. 2011; Gjoka et al. 2010], Bio-Informatics [Plake et al. 2006] and so on. The importance of Web Data Extraction systems depends on the fact that, today, a large (and quickly growing) amount of information is continuously produced, shared and consumed online: Web Data Extraction systems allow to eciently collect this information with a limited human eort. The availability and analysis of collected data is an indefeasible requirement to understand complex social, scientic and economic phenomena which generated the information itself. So, for instance, collecting digital traces produced by human users in Social Web platforms like Facebook, YouTube or Flickr is the key step to verify sociological theories on a large scale [Kleinberg 2000; Backstrom et al. 2011] or to check whether mathematical models developed in the eld of Complex Networks [Newman 2003] are able to correctly explain human behaviors. In the commercial eld, the Web provides a wealth of public domain information which can be retrieved, for instance, from the Web page of a rm or from an ecommerce platform. A rm can probe the Web to acquire and analyze information about the activities of its competitors; such a process is generally known as Competitive Intelligence [Chen et al. 2002; Zanasi 1998] and it is a success factor for the rm: the rm, in fact, is able to quickly identify the opportunities provided by the market, to anticipate the activities of its competitors as well as to learn from their faults and successes. 1.1 Challenges of Web Data Extraction techniques

The design and implementation of Web Data Extraction Systems has been considered from dierent perspectives and it leverages on scientic results and tools coming from various disciplines like Machine Learning, Logic and Natural Language Processing. In the design of a Web Data Extraction system, many factors must be taken into account; some of these factors are independent of the specic application domain in which we plan to perform Web Data Extraction. Other factors, instead, heavily depend on the particular features of the application domain: as a consequence, some technological solutions which appear to be eective in some application contexts are not suitable in other ones. To better clarify this concept, observe that some approaches focus on static HTML Web pages and use the tags composing a page
ACM Computing Surveys, Vol. V, No. N, July 2012.

Emilio Ferrara et al.

along with their hierarchical organization to extract information. These approaches are able to achieve a high level of accuracy as well as to eciently scale over large document collections, but, unfortunately, the results they obtain are no longer replicable if other kind of Web sources are considered (e.g., Web pages dynamically generated after lling a Web form). Such a reasoning tells us that, in its most general formulation, the problem of extracting data from the Web is hard because it is constrained by several requirements. The key challenges we can encounter in the design of a Web Data Extraction system can be summarized as follows: Web Data Extraction techniques often require the help of human experts. A rst challenge consists of providing a high degree of automation by reducing human eorts as much as possible. Human feedbacks, however, may play an important role in raising the level of accuracy achieved by a Web Data Extraction system. A related challenge is, therefore, to identify a reasonable trade-o between the need of building highly automated Web Data Extraction procedures and the requirement of achieving highly accurate performance. Web Data Extraction techniques should be able to process large volumes of data in relatively short time. Such a need is particularly urgent in the eld of Business and Competitive Intelligence because a rm needs to perform timely analysis of market conditions. Applications in the eld of Social Web or, more in general, applications dealing with human related data must provide solid privacy guarantees. Therefore, potential (even if unintentional) attempts to violate user privacy should be timely and adequately identied and advertised. Approaches relying on Machine Learning often require a signicantly large training set of manually labeled Web pages. In general, the task of labeling pages is time-expensive and error-prone and, therefore, in many cases we can not assume the existence of labeled pages. In some cases, a Web Data Extraction tool has to routinely extract data from a Web Data source which can evolve over time. Web sources are continuously evolving and structural changes happen with no forewarning, thus are unpredictable. Eventually, in real-world scenarios it emerges the need of maintaining these systems, that might stop working correctly if lacking of exibility to detect and face structural modications of related Web sources. 1.2 Related works

The theme of Web Data Extraction is covered by a number of reviews. Laender et al. [Laender et al. 2002], presented a survey that oers a rigorous taxonomy to classify Web Data Extraction systems. They introduced a set of criteria and a qualitative analysis of various Web Data Extraction tools. Kushmerick [Kushmerick 2002] tracked a prole of nite-state approaches to the Web Data Extraction problem. The author analyzed both wrapper induction approaches (i.e., approaches capable of automatically generating wrappers by exploiting suitable examples) and maintenance ones (i.e., the update of a wrapper each time the structure of the Web source changes). In [Kushmerick 2002], Web Data ExACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

traction techniques derived from Natural Language Processing and Hidden Markov Models were also discussed. On the wrapper induction problem, Flesca et al. [Flesca et al. 2004] and Kaiser and Miksch [Kaiser and Miksch 2005] surveyed approaches, techniques and tools. The latter paper, in particular, provided a model describing the architecture of an Information Extraction system. Chang et al. [Chang et al. 2006] introduced a tri-dimensional categorization of Web Data Extraction systems, based on task diculties, techniques used and degree of automation. Fiumara [Fiumara 2007] applied these criteria to classify four among the state-of-the-art Web Data Extraction systems as of 2007. A relevant work on Information Extraction is due to Sarawagi [Sarawagi 2008] and, in our opinion, anybody who intends to approach this discipline should read it. To the best of our knowledge, the latest work from Baumgartner et al. [Baumgartner et al. 2009] is the most updated survey on the state-of-the-art of the discipline as of this date. 1.3 Our contribution

The goal of this survey is to provide a structured and comprehensive overview of the research eorts made in the eld of Web Data Extraction and as well as to provide an overview of most recent results in the literature. In addition, we try to adopt a point of view dierent with respect to that used in other survey on this discipline: most papers, in fact, presented a list of tools, reporting a feature-based classication or an experimental comparison of these tools. Many of these papers are solid starting points in the study of this area. Unlike the existing surveys, our ambition is to provide a classication of existing Web Data Extraction techniques in terms of the application domains in which they have been employed. We want to shed light on the various research directions made in this eld as well as to understand to what extent techniques initially applied in a particular application domain have been later re-used in others. To the best of our knowledge, this is the rst survey deeply analyzing time a Web Data Extraction system from a perspective of their application elds. Although, we also provide a detailed discussion of techniques to perform Web Data Extraction. We identify two main categories, i.e., approaches based on Tree Matching algorithms and approaches based on Machine Learning algorithms. For each category, we rst describe the basic techniques employed in that category and after this we illustrate their variants. We show also how each category addresses the problems of automatic wrapper generation and maintenance. After that, we focus on applications that are strictly interconnected with Web Data Extraction tasks. We want to cover in particular enterprise, social and scientic applications and discover which elds have already been approached (e.g., advertising engineering, enterprise solutions, Business and Competitive intelligence, etc.) and which are going to be, looking ahead the immediate future (e.g., Bio-informatics, Web Harvesting, etc.). We also discuss about the potential of cross-fertilization, i.e., whether strategies employed in a given domain can be re-used in others or, vice versa, if some applications can be adopted only in particular application domains.
ACM Computing Surveys, Vol. V, No. N, July 2012.

Emilio Ferrara et al.

1.4

Organization of the survey

This survey is organized into two main parts. The rst part is devoted to provide some general denitions which are helpful to understand the material proposed in the survey. To this purpose, Section 2 illustrates the techniques exploited for collecting data from Web sources and the algorithms that underly most of Web Data Extraction systems. The main features of existing Web Data Extraction systems are largely discussed in Section 3. The second part of this work is about the applications of Web Data Extraction systems to real-world scenarios. In detail, in Section 4 we identied two main domains in which Web Data Extraction techniques have been employed, i.e., applications at the enterprise level and at the Social Web level. Application at the enterprise level are described in Section 4.1, whereas applications at the Social Web level are covered in Section 4.2. This part concludes discussing the opportunities of cross-fertilization among dierent application scenarios (see Section 4.3). Finally, in Section 5 we draw our conclusions and discuss novel applications of Web Data Extraction techniques that might arise in the future. 2. TECHNIQUES

The rst part of this survey is dedicated to the discussion of the techniques adopted in the eld of the Web Data Extraction. The rst trials to extract data from the Web are dated back in early nineties, as reported by [Kaiser and Miksch 2005; Sarawagi 2008]. In the early stage, this discipline borrowed approaches and techniques from Information Extraction (IE) literature. In particular, two classes of strategies emerge [Kaiser and Miksch 2005]: learning techniques and knowledge engineering techniques also called as learning-based and rule-based approaches, respectively [Sarawagi 2008]. These classes share a common rationale: the former was thought to develop systems that require human expertise to dene rules (for example, regular expressions or program snippets) to successfully accomplish the data extraction. These approaches require specic domain expertise: users that design and implement the rules and train the system must have programming experience and a good knowledge of the domain in which the data extraction system will operate and, possibly, the ability of envisaging potential usage scenarios and tasks assigned to the system. On the other hand, also some approaches of the latter class involve strong familiarity with both the requirements and the functions of the platform, so the human engagement is essential. To reduce the commitment of human domain experts, a number of strategies have been devised. Some of them have been developed in the context of Articial Intelligence literature, by means of the adoption of specic algorithms using the structure of Web pages to identify and extract data. Some others are borrowed from the Machine Learning discipline, thought as supervised or semi-supervised learning techniques which allow the design of systems capable of being trained by examples and then to autonomously extract data from similar (or even dierent) domains. In the following we will discuss separately these two research lines. In detail, Section 2.1 presents those strategy based on the denition of algorithms capable
ACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

of identifying information exploiting the semi-structured nature of Web pages. In Section 2.2 we introduce the concept of Web wrappers and explain how the techniques discussed in Section 2.1 are incorporated into Web wrappers, and discuss techniques for their generation and maintenance. Section 2.3, instead, poses the attention on Machine Learning strategies relying on the possibility of learning from labeled examples to extract information from not-previously seen Web pages. 2.1 Tree-based techniques

One of the most exploited features in Web Data Extraction is the semi-structured nature of Web pages. These can be naturally represented as labeled ordered rooted trees, where labels represent the tags proper of the HTML mark-up language syntax, and the tree hierarchy represents the dierent levels of nesting of elements constituting the Web page. The representation of a Web page by using a labeled ordered rooted tree is usually referred as DOM (Document Object Model), whose detailed explanation is out of the scope of this survey but has been largely regulated by the World Wide Web Consortium1 . The general idea behind the Document Object Model is that HTML Web pages are represented by means of plain text, which contains HTML tags, i.e., particular keywords dened in the mark-up language that can be interpreted by the browser to represent the elements specic of a Web page (e.g., hyper-links, buttons, images and so forth), so as free-text. HTML tags may be nested one into another, forming a hierarchical structure. This hierarchy is captured in the DOM by the document tree, whose nodes represent HTML tags. The document tree (henceforth also referred as DOM tree) has been successfully exploited for Web Data Extraction purposes in a number of techniques discussed in the following. 2.1.1 Addressing elements in the document tree: XPath. One of the main advantage of the adoption of the Document Object Model for the HTML language is the possibility of exploiting some tools typical of XML languages (and HTML is to all eects a dialect of the XML). In particular, the XML Path Language (or, briey, XPath) provides with a powerful syntax to address specic elements of a XML document (and, to the same extent, of HTML Web pages) in a simple manner. XPath has been dened by the World Wide Web Consortium2 , so as DOM. Although describing the syntax of XPath is not the core argument of this section, we provide Figure 1 as an example to explain how XPath can be used to address elements of a Web pages. There exist two possible ways to use XPath: (i) to identify a single element in the document tree, or (ii) to address multiple occurrences of the same element. In the former case, illustrated in Figure 1(A), the dened XPath identies just a single element on the Web page (namely, a table cell); in the latter, showed in Figure 1(B), the XPath identies multiple instances of the same type of element (still a table cell) sharing the same hierarchical location. To the purpose of Web Data Extraction, the possibility of exploiting a such powerful tool has been of utmost importance and the adoption of XPath as the tool to address elements in Web page has been largely exploited in the literature. The
1 For 2 For

further details consider the following link: http://www.w3.org/DOM/ the specications see: http://www.w3.org/TR/xpath/
ACM Computing Surveys, Vol. V, No. N, July 2012.

Emilio Ferrara et al.

(A) /html[1]/body[1]/table[1]/tr[1]/td[1] html (B) /html[1]/body[1]/table[1]/tr[2]/td html

head

body table

head

body table

tr td . . . td

tr td td td . . .

tr td . . .

tr td . . . td

tr td td td . . .

tr td . . .

Fig. 1.

Example of XPath(s) on the document tree, selecting one (A) or multiple (B) items.

major weakness of XPath is related to its lack of exibility: each XPath expression is strictly related to the structure of the Web page on top of which it has been dened (this limitation has been partially mitigated by the denition of relative path expressions instead of only absolute ones, in latest releases, e.g., XPath 2.03 ). This implies that even minor changes applied to the structure of a Web page might corrupt the correct functioning of a XPath expression dened on the previous version of the page itself. The emergence of such a problem let us introduce some techniques, usually referred as tree matching strategies, devoted to assess and measure the similarity among dierent documents, so as to identify structural variations in their DOM trees. These techniques, discussed in the following, have been exploited in the literature to overcome dierent situations in which exibility in the identication of elements within Web pages was a requirement to make a Web Data Extraction platform robust and reliable. 2.1.2 Tree edit distance matching algorithms. The rst technique we describe is called tree edit distance matching. The problem of computing the tree edit distance between trees is a variation on the theme of the classic string edit distance problem. Given two labeled ordered rooted trees A and B, the problem is to nding a matching to transform A in B (or the vice-versa) with the minimum number of operations. The set of possible operations consists of node deletion, insertion or replacement. At each operation might be applied a cost, and in that case, the task turns in a cost-minimization problem (i.e., nding the sequence of operations of minimum cost to transform A in B). The reasoning above is formally encoded in the denition of mapping, borrowed by [Liu 2011]. A mapping M between two trees A and B is dened as a set of ordered pairs (i, j), one from each tree, satisfying the following conditions (i1 , j1 ), (i2 , j2 ) M . (1) i1 = i2 if and only if j1 = j2 (2) A[i1 ] is on the left of A[i2 ] if and only if B[i1 ] is on the left of B[i2 ] (3) A[i1 ] is an ancestor of A[i2 ] if and only if B[i1 ] is an ancestor of B[i2 ] where with the notation A[ix ] we indicate the x-th node of the tree A in a preorder visit of the tree.
3 http://www.w3.org/TR/xpath20/

ACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

A number of intuitive consequences emerge from this denition of mapping: Each node must appear no more than once in a mapping; The order among siblings is preserved; The hierarchical relations among nodes is unchanged. A number of techniques to approach this problem have been proposed [Wang et al. 1998; Chen 2001], all providing the support to all three types of operations on nodes (i.e., node deletion, insertion and replacement) but plagued by high computational costs. In addition, in [Zhang et al. 1992] it has been proved that the formulation for non ordered trees is NP-complete. The simple tree matching algorithm. A computationally ecient solution for the problem of the tree edit distance matching is provided by the algorithm called simple tree matching [Selkow 1977], and its variants. This optimized strategy comes at a cost: node replacement is not an allowed operation during the matching procedure the shortcomings of this aspects will be further discussed below. The pseudo-code of the simple tree matching algorithm is provided in Algorithm 1, that adopts the following notation: d(n) represents the degree of a node n (i.e., the number of rst-level children); T (i) is the i-th subtree of the tree rooted at node T. Algorithm 1 SimpleTreeMatching(T , T )
1: 2: 3: 4: 5: 6: 7: 8: 9: 10: 11: 12: 13: 14: 15: 16: 17: 18:

if T has the same label of T then m d(T ) n d(T ) for i = 0 to m do M [i][0] 0; end for for j = 0 to n do M [0][j] 0; end for for all i such that 1 i m do for all j such that 1 j n do M [i][j] Max(M [i][j 1], M [i 1][j], M [i 1][j 1] + W [i][j]) where W [i][j] = SimpleTreeMatching(T (i 1), T (j 1)) end for end for return M[m][n]+1 else return 0 end if

The computational cost of the simple tree matching is O(nodes(A) nodes(B)), where nodes(T ) is the function that returns the number of nodes in a tree (or a sub-tree) T ; the low cost ensures excellent performance when applied to HTML trees, that might be complex and rich of nodes.
ACM Computing Surveys, Vol. V, No. N, July 2012.

10

Emilio Ferrara et al.

There are two main limitations of this algorithm: It can not match permutation of nodes; No level crossing is allowed (i.e., it is not possible to match at dierent hierarchical levels). Despite these intrinsic limits, this technique appears to t very well to the purpose of matching HTML trees in the context of Web Data Extraction systems. In fact, it has been adopted in several scenarios [Kim et al. 2007; Zhai and Liu 2005; Zhang et al. 2010; Ferrara and Baumgartner 2011a; 2011b; 2011c]. The reasons for its choice are dierent; for example, the simple tree matching algorithm evaluates similarity between two trees by producing the maximum matching through dynamic programming, without computing inserting, relabeling and deleting operations; moreover, approximate tree edit distance algorithms relies on complex implementations to achieve good performance, instead simple tree matching, and its variants, are easy to implement. A normalized variant of the simple tree matching is called normalized simple tree matching, whose result is given by: N ST M (A, B) = SimpleT reeM atching(A, B) (nodes(A) + nodes(B))/2

The normalization is computed considering the average size of nodes in the given trees (or sub-trees). The weighted tree matching algorithm. Another variant of the simple tree matching is discussed in the following, and is called weighted tree matching. It adjusts the similarity values provided by the original simple tree matching by introducing a re-normalization factor. The weighted tree matching, whose pseudo-codify is reported in Algorithm 2, has been recently presented in [Ferrara and Baumgartner 2011a]. Algorithm 2 WeightedTreeMatching(T , T ) 1: {Change line 11 with the following code} 2: if m > 0 AND n > 0 then 3: return M[m][n] * 1 / Max(t(T ), t(T )) 4: else 5: return M[m][n] + 1 / Max(t(T ), t(T )) 6: end if In Algorithm 2, the notation t(n) represents the number of total siblings of a node n including itself. Note that Algorithm 2 reports the dierential version with respect to the simple tree matching described in Algorithm 1. The advantage of the weighted tree matching is that it better reects a measure of similarity between two trees. In fact, in the simple tree matching algorithm the assigned matching value is always equal to one. Instead, the weighted tree matching algorithm assumes that less importance (i.e., a lower weight) is assigned to changes in the structure
ACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

11

of the tree, when they occur in deeper sub-levels. This kind of changes can be, for example, missing or added leaves, truncated or added branches, etc. Also, a lower weight is accounted when changes occur in sub-levels with many nodes. In the weighted tree matching, the weighted value assigned to a match between two nodes is equal to one divided by the greatest number of siblings with respect to the two compared nodes, including themselves. This helps reducing the impact of missing/added nodes on the nal similarity value. Moreover, before assigning a weight, the algorithm checks if it is comparing two leaves or a leaf with a node which has children (or two nodes which both have children). The nal contribution of a sub-level of leaves is the sum of assigned weighted values to each leaf (cfr. Code Line (4,5)); thus, the contribution of the parent node of those leaves is equal to its weighted value multiplied by the sum of contributions of its children (cfr. Code Line (2,3)). For each sub-level of leaves, the maximum sum of assigned values is 1; thus, for each parent node of that sublevel, the maximum value of the multiplication of its contribution with the sum of contributions of its children is 1; each sub-tree, individually considered, contributes with a maximum value of 1. In the last recursion of this bottom-up algorithm, the two roots will be evaluated. The resulting value at the end of the process is the measure of similarity between the two trees, expressed in the interval [0,1]. The closer to 1 the nal value, the more similar the two trees. Let us analyze the behavior of the algorithm with an example often used in the literature [Yang 1991; Zhai and Liu 2005; Ferrara and Baumgartner 2011a] to explain the simple tree matching (see Figure 2). In that gure, A and B are two very simple generic rooted labeled trees (i.e., the same structure of HTML document trees). They show several similarities except for some missing nodes/branches. By applying the weighted tree matching algorithm, in the rst step the contributions assigned to leaves reect the considerations discussed above: for example, a value of 1 is established for nodes (h), (i) and (j) belonging to A, although two 3 of them are missing in B. Going up to parents, the summation of contributions of matching leaves is multiplied by the relative value of each node (e.g., in the rst 1 sub-level, the contribution of each node is 4 because of the four rst-sublevel nodes in A). Once completed these operations for all nodes of the sub-level, values are added and the nal measure of similarity for the two trees is obtained. Intuitively, in more complex and deeper trees, this process is iteratively executed for all the sub-levels. The deeper a mismatch is found, the less its missing contribution will aect the nal measure of similarity. Analogous considerations hold for missing/added nodes and branches, sub-levels with many nodes, etc. Table I shows the two matrices adopted by the weighted tree matching algorithm, M and W that contain the matching and the weights at each iteration.
W N6 N7 N18 1 2 0 N19 0 1 2 M 0 N6 N6-7 0 0 0 0 N18 0 1 2 1 2 N18-19 0 1 2 1 W N9 N10 N18 0 1 2 N19 1 2 0 M 0 N9 N9-10 0 0 0 0 N18 0 0 1 2 N18-19 0 1 2 1 2

In this example, the weighted tree matching between A and B returns a measure 3 of similarity of 8 (0.375) whereas the simple tree matching would return a mapping
ACM Computing Surveys, Vol. V, No. N, July 2012.

12
(A)

Emilio Ferrara et al.

a N1 (B) a N15

b N2 1 4 (N6+N7) d N6 1 2 e N7 1 2

c N3 1 N8 4 f N8 1 2

b N4 1 4 (N9+N10) e N9 1 2 d N10 1 2

c N5 1 N11 4 g N11 1 2 (N12+N13+N14)

b N16 1 4 (N18+N19) d N18 1 2 e N19 1 2

c N17 1 4 (N20+N21) g N20 1 N22 2 h N22 1 3 f N21 1 2

h N12 1 3

i N13 1 3

j N14 1 3

Fig. 2. Example of application of the weighted tree matching algorithm for the comparison of two labeled rooted trees, A and B.
W N20 N21 N8 0 1 2 M 0 N20 N20-21 0 0 0 0 N8 0 0 1 2 M 0 N22 0 0 0 N12 0 1 3 N12-13 0 1 3 N12-14 0 1 3

W N22

N12 1 3

N13 0

N14 0

W N20 N21

N11 1 6 0

M 0 N20 N20-21

0 0 0 0

N11 0 1 6 1 6

W N16 N17

N2 1 4 0

N3 0 1 8

N4 1 8 0

N5 0 1 24

M 0 N16 N16-17

0 0 0 0

N2 0 1 4 1 4

N2-3 0 1 4 3 8

N2-4 0 1 4 3 8

N2-5 0 1 4 3 8

Table I.

W and M matrices for each matching subtree.

value of 7. The main dierence on results provided by these two algorithms is the following: the weighted tree matching intrinsically produces a proper measure of similarity between the two compared trees while the simple tree matching returns the mapping value. A remarkable feature of the weighted tree matching algorithm is that, the more the structure of considered trees is complex and similar, the more the measure of similarity will be accurate. On the other hand, for simple and quite dierent trees the accuracy of this approach is lower than the one ensured by the simple tree matching. 2.2 Web wrappers

In the previous section we discussed some algorithms that might be adopted to identify information exploiting the semi-structured format of HTML documents. In the following we will discuss those procedures that might adopt the techniques presented above to carry out the data extraction. In the literature, any procedure that aims at extracting structure data from unstructured (or semi-structured) data sources is usually referred as wrapper. In the context of Web Data Extraction we provide the following denition: Denition 2.1 Web wrapper. A procedure, that might implement one or dierent classes of algorithms, which seeks and nds data required by a human user, extracting them from unstructured (or semi-structured) Web sources, and transforming them into structured data, merging and unifying this information for further processing, in a semi-automatic or fully automatic way.
ACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

13

Web wrappers are characterized by a life-cycle constituted by several steps: (1) Wrapper generation: the wrapper is dened according to some technique(s); (2) Wrapper execution: the wrapper runs and extracts information continuously; (3) Wrapper maintenance: the structure of data sources may change and the wrapper should be adapted accordingly to keep working properly. In the remainder of this section we will discuss these three dierent phases. In particular, the rst two steps of a wrapper life-cycle, its generation and execution, are discussed in Section 2.2.1; these steps might be implemented manually, for example by dening and executing regular expressions over the HTML documents; alternatively, which is the aim of Web Data Extraction systems, wrappers might be dened and executed by using an inductive approach a process commonly known as wrapper induction [Kushmerick 1997]. Web wrapper induction is challenging because it requires high level automation strategies. There exist also hybrid approaches that make possible for users to generate and run wrappers semi-automatically by means of visual interfaces. The last step of a wrapper life-cycle is the maintenance: Web pages change their structure continuously and without forewarning. This might corrupt the correct functioning of a Web wrapper, whose denition is usually tightly bound to the structure of the Web pages adopted for its generation. Dening automatic strategies for wrapper maintenance is of outstanding importance to guarantee correctness of extracted data and robustness of Web Data Extraction platforms. Some methodologies have been recently presented in the literature, and are discussed in Section 2.2.2. 2.2.1 Wrapper generation and execution. The rst step in wrappers life-cycle is their generation. Early Web Data Extraction platforms provided only support for manual generation of wrappers, which required human expertise and skills in programming languages to write scripts able to identify and extract selected pieces of information within a Web page. In late nineties they made their appearance more advanced Web Data Extraction systems. The core feature provided was the possibility for their users to dene and execute Web wrappers by means of interactive graphical users interfaces (GUIs). In most cases, it was not required any deep understanding of a wrapper programming language, as wrappers were generated automatically (or semi-automatically) by the system exploiting directives given by users by means of the platform interface. In the following we discuss in detail three types of rationales underlying these kind of platforms, namely regular expressions, wrapper programming languages and tree-based approaches. The details regarding the features and of dierent Web Data Extraction platforms, instead, will be described in great detail in Section 3. Regular-expression-based approach. One of the most common approaches is based on regular expressions, which are a powerful formal language used to identify strings or patterns in unstructured text on the basis of some matching criteria. Rules could be complex so, writing them manually, could require much time and a great expertise: wrappers based on regular expressions dynamically generate rules to extract
ACM Computing Surveys, Vol. V, No. N, July 2012.

14

Emilio Ferrara et al.

desired data from Web pages. Usually, writing regular expressions on HTML pages relies on the following criteria: word boundaries, HTML tags, tables structure, etc. The advantage of platforms relying on regular expressions is that the user can usually select (for example by means of a graphical interface) one (or multiple) element(s) in a Web page, and the system is able to automatically infer the appropriate regular expression to identify that element in the page. Then, a wrapper might be created so that to extract similar elements, from other Web pages with the same structure of the one adopted to infer the regular expressions. A notable tool implementing regular-expression-based extraction is W4F [Sahuguet and Azavant 1999]. W4F adopts an annotation approach: instead of challenging users to deal with the HTML documents syntax, W4F eases the design of the wrapper by means of a wizard procedure. This wizard allows users to select and annotate elements directly on the Web page. W4F produces the regular expression extraction rules of the annotated items and provides them to users. A further step, which is the optimization of the regular expressions generated by W4F, is delegated to expert users in fact, the tool is not always able to provide the best extraction rule. By fully exploiting the power of regular expressions, W4F extraction rules include match and also split expressions, which separates words, annotating dierent elements on the same string. The drawback of the adoption of regular expressions is their lack of exibility. For example, whenever even a minor change occurs in the structure or content of a Web page, each regular expression is very likely to stop working, and must be rewritten. This process implies a big commitment by human users, in particular in the maintenance of systems based on regular expressions. For this reasons more exible and powerful languages have been developed to empower the capabilities of Web Data Extraction platforms. Logic-based approach. One example of powerful languages developed for data extraction purposes comes from the Web specic wrapper programming languages. Tools based on wrapper programming languages consider Web pages not as simply text strings but as semi-structured tree documents, whereas the DOM of the Web page represents its structure where nodes are elements characterized by both their properties and their content. The advantage of such an approach is that wrapper programming languages might be dened to fully exploit both the semi-structured nature of the Web pages and their contents the former aspect lacks in regularexpression-based systems. The rst powerful wrapping language has been formalized by Gottlob and Koch [Gottlob and Koch 2004a]. The information extraction functions implemented by this wrapping language rely on monadic datalogs over trees [Gottlob and Koch 2004b]. The authors demonstrated that monadic datalogs over tree are equivalent to monadic second order logic (MSO), and hence very expressive. However, unlike MSO, a wrapper in monadic datalogs can be modeled nicely in a visual and interactive step-by-step manner. This makes this wrappring language suitable for being incorporated into visual tools, satisfying the condition that all its constructs can be implemented through corresponding visual primitives. A bit of avor on the functioning of the wrapping language is provided in the following. Starting from the unranked labeled tree representing the DOM of the Web page, the algorithm re-labels nodes, truncates the irrelevant ones, and nally
ACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

15

returns a subset of original tree nodes, representing the selected data extracted. The rst implementation of this wrapping language in a real-world scenarios is due to Baumgartner et al. [Baumgartner et al. 2001b; 2001a]. They developed the Elog wrapping language that implements almost all monadic datalog information extraction functions, with some minor restrictions. The Elog language is used as the core extraction method of the Lixto Visual Wrapper system. This platform provides a GUI to select, through visual specication, patterns in Web pages, in hierarchical order, highlighting elements of the document and specifying relationships among them. Information identied in this way could be too general, thus the system allows users to add some restricting conditions, for example before/after, not-before/notafter, internal and range conditions. Finally, selected data are translated into XML documents, by using pattern names as XML element names, obtaining structured data from semi-structured Web pages. Tree-based approach [partial tree alignment]. The last technique discussed in this part relates to wrapper generation and is called partial tree alignment. It has been recently formalized by Zhai and Liu [Zhai and Liu 2005; 2006] and the authors also developed a Web Data Extraction system based on it. This technique relies on the idea that information in Web documents usually are collected in contiguous regions of the page, called data record regions. The strategy of partial tree alignment consists in identifying and extracting these regions. In particular, the authors take inspiration from tree matching algorithms, by using the already discussed tree edit distance matching (see Section 2.1.2). The algorithm works in two steps: (1) Segmentation; (2) Partial tree alignment. In the rst phase, the Web page is split in segments, without extracting any data. This pre-processing phase is instrumental to the latter step. In fact, the system not only performs an analysis of the Web page document based on the DOM tree, but also relies on visual cues (like in the spatial reasoning technique, see Section 2.3.1), trying to identify gaps between data records. This step is useful also because helps the process of extracting structural information from the HTML document, in that situations when the HTML syntax is abused, for example by using tabular structure instead of CSS to arrange the graphical aspect of the page. In the second step, the partial tree alignment algorithm is applied to data records earlier identied. Each data record is extracted from its DOM sub-tree position, constituting the root of a new single tree. This, because each data record could be contained in more than one non-contiguous sub-tree in the original DOM tree. The partial tree alignment approach implies the alignment of data elds with certainty, excluding those that can not be aligned, to ensure a high degree of precision. During this process no data items are involved, because partial tree alignment works only on tree tags matching, represented as the minimum cost, in terms of operations (i.e., node removal, node insertion, node replacement), to transform one node into another one. The drawback of this characteristic of the algorithm is that its recall performance (i.e., the ability of recovering all expected information) might decay in case of complex HTML document structures. In addition, also in the case of the partial tree alignment, the functioning of this strategy is strictly related with the
ACM Computing Surveys, Vol. V, No. N, July 2012.

16

Emilio Ferrara et al.

structure of the Web page at the time of the denition of the alignment. This implies that the method is very sensitive even to small changes, that might compromise the functioning of the algorithm and the correct extraction of information. Thus, even in this approach, the problem of the maintenance arises with outstanding importance. 2.2.2 The problem of wrapper maintenance. Wrapper generation, regardless the adopted technique, is one aspect of the problem of data extraction from Web sources. On the other hand, wrapper maintenance is equally important, so that Web Data Extraction platform may reach high levels of robustness and reliability, hand in hand with the level of automation and low level of human engagement. In fact, dierently from static documents, Web pages dynamically change and evolve, and their structure may change, sometimes with the consequence that previously dened wrappers are no longer able to successfully extract data. In the light of these assumptions, one could argument that wrapper maintenance is a critical step of the Web Data Extraction process. Even though, this aspect has not acquired lot of attention in the literature (much less than the problem of wrapper generation), unless latest years. In the early stage, in fact, wrapper maintenance was performed manually: users that usually design Web wrappers, were updating or rewriting these wrappers every time the structure of a given Web page was modied. The manual maintenance approach was tting well for small problems, but becomes unsuitable if the pool of Web pages largely increases. Since in the enterprise scenarios regular data extraction tasks might involve thousand (or even more) Web pages, dynamically generated and frequently updated, the manual wrapper maintenance is not anymore a feasible solution for real-world applications. For these reasons, the problem of automatizing the wrapper maintenance has been faced in recent literature. For example, the rst eort in the direction of automatic wrapper maintenance has been presented by Kushmerick [Kushmerick 2002], who dened for rst the concept of wrapper verication. The task of wrapper verication arises as a required step during wrapper execution, in which a Web Data Extraction system assess if dened Web wrappers work properly or, alternatively, their functioning is corrupted due to modications to the structure of underlying pages. Subsequently, the author discussed some techniques of semi-automatic wrapper maintenance, to handle simple problems. The rst method that tries to automatize the process of wrapper maintenance has been developed by Meng et al. [Meng et al. 2003], and it is called schemaguided wrapper maintenance. It relies on the denition of XML schemes during the phase of the wrapper generation, to be exploited for the maintenance during the execution step. More recently, Ferrara and Baumgartner [Ferrara and Baumgartner 2011a; 2011b; 2011c] developed a system of automatic wrapper adaptation (a kind of maintenance that occur to modify Web wrappers according to the new structure of the Web pages) relying on the analysis of structural similarities between dierent versions of the same Web page using a tree-edit algorithm. In the following, we discuss these two strategies of wrapper maintenance. Schema-guided wrapper maintenance. The rst attempt to deal with the problem of wrapper maintenance providing a high level of automation has been presented
ACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

17

by Meng et al. [Meng et al. 2003]. The authors developed SG-WRAM (SchemaGuided WRApper Maintenance), a strategy for Web Data Extraction that is built on top of the assumption, based on empirical observations, that changes in Web pages, even substantial, often preserve: Syntactic features: syntactic characteristics of data items like data patterns, string lengths, etc., are mostly preserved. Hyperlinks: HTML documents are often enriched by hyperlinks that are seldom removed in subsequent modications of the Web pages. Annotations: descriptive information representing the semantic meaning of a piece of information in its context is usually maintained. In the light of these assumptions, the authors developed a Web Data Extraction system that, during the phase of wrapper generation, creates schemes which will be exploited during the phase of wrapper maintenance. In detail, during the generation of wrappers, the user provides HTML documents and XML schemes, specifying a mapping between them. Later, the system will generate extraction rules and then it will execute the wrappers to extract data, building an XML document according to the specied XML schema. During the wrapper execution phase, an additional component is introduced in the pipeline of the data extraction process: the wrapper maintainer. The wrapper maintainer checks for any potential extraction issue and provides an automatic repairing protocol for wrappers which fail their extraction task because of modications in the structure of related Web pages. The repairing protocol might be successful, and in that case the data extraction continues, or it might fail in that case warnings and notications rise. The XML schemes are dened in the format of a DTD (Document Type Denition) and the HTML documents are represented as DOM trees, according to what explained in Section 2.1. The SG-WRAM system builds corresponding mappings between them and generates extraction rules in the format of XQuery expressions4 . Automatic wrapper adaptation. Another strategy for the automatic maintenance of Web wrappers has been recently presented by Ferrara and Baumgartner [Ferrara and Baumgartner 2011a; 2011b; 2011c]. In detail, it is a method of automatic wrapper adaptation that relies on the idea of comparing helpful structural information stored in the Web wrapper dened on the original version of the Web page, searching for similarities in the new version of the page, after any structural modication occurs. The strategy works for dierent techniques of data extraction implemented by the wrapping system. For example, it has been tested by using both XPath (see Section 2.1.1) and the Elog wrapping language (see Section 2.2.1). In this strategy, elements are identied and represented as sub-trees of the DOM tree of the Web page, and can be exploited to nd similarities between two dierent versions of the same document. We discuss an example adopting XPath to address a single element in a Web page, as reported in the example in Figure 1(A). The rationale behind the automatic wrapper adaptation is to search for some elements, in the modied version of the
4 For

XQuery specications see: http://www.w3.org/TR/xquery/

ACM Computing Surveys, Vol. V, No. N, July 2012.

18

Emilio Ferrara et al.

Web page, that share structural similarities with the original one. The evaluation of the similarity is done on the basis of comparable features (e.g., subtrees, attributes, etc.). These elements are called candidates: among them, the one showing the higher degree of similarity with the element in the original page, is matched in the new version of the page. The algorithm adopted to compute the matching among the DOM trees of the two HTML documents is the weighted tree matching, already presented in Section 2.1.2. Further heuristics are adopted to assess the similarity of nodes, for example exploiting additional features exhibited by DOM elements. In some cases, for example, their textual contents might be compared, according to string distance measures such as Jaro-Winkler [Winkler 1999] or bigrams [Collins 1996], to take into account also the content similarity of two given nodes. It is possible to extend the same approach in the case in which the XPath identies multiple similar elements on the original page (e.g., an XPath selecting results of a search in a retail online shop, represented as table rows, divs or list items), as reported in Figure 1(B). In detail, it is possible to identify multiple elements sharing a similar structure in the new page, within a custom level of accuracy (e.g., establishing a threshold value of similarity). The authors implemented this approach in a commercial tool Lixto, describing how the pipeline of the wrapper generation has been modied to allow the wrappers to automatically detect and adapt their functioning to structural modications occurring to Web pages [Ferrara and Baumgartner 2011c]. In this context, the strategy proposed was to acquire structural information about those elements the original wrapper extracts, storing them directly inside the wrapper itself. This is done, for example, generating signatures representing the DOM sub-tree of extracted elements from the original Web page, stored as a tree diagram, or as a simple XML documents. During the execution of Web wrappers, if any extraction issue occurs due to a structural modication of the page, the wrapper adaptation algorithm automatically starts and tries to adapt the wrapper to the new structure. 2.3 Machine Learning approaches

Not only the literature in the context of Articial Intelligence and Algorithms focused on the problem Web Data Extraction, but also the eld of Machine Learning has proposed several interesting solutions. In particular, Machine Learning techniques t well to the purpose of extracting domain-specic information from Web sources, since they rely on training sessions during which a system acquires a domain expertise. On the other hand, training learning-based Web Data Extraction systems usually requires large amount of manually labeled Web pages, that implies a bootstrap period in which the systems require a high level of human engagement. Moreover, during the manual labeling phase, domain experts should provide both positive and negative examples, acquired from dierent websites but also in the same website. Particular attention should be paid to providing examples of Web pages belonging to the same domain but exhibiting dierent structures. This, because, even in the same domain scenario, templates usually adopted to generate dynamic contents Web pages, dier, and the system should be capable of learning how to extract information in these contexts. Statistical Machine Learning systems were also developed, relying on conditional models [Phan et al. 2005] or
ACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

19

adaptive search [Turmo et al. 2006] as an alternative solution to human knowledge and interaction. WIEN. In the early literature, some interesting wrapper induction systems have been presented. In fact, Kushmerick developed the rst wrapper induction system, called WIEN [Kushmerick 2000], based on dierent inductive learning techniques. Interestingly, WIEN was capable of automatically labeling training pages, representing de facto a hybrid system whose training process was accelerated and implied low human engagement. The ip side of the high automation of WIEN was the big number of limitations related to its inferential system: for example, the data extraction process was not capable of dealing with missing values a case that occurs on a frequent base and posed serious limitations on the adaptability of WIEN to real-world scenarios. SoftMealy. Another learning-based Web Data Extraction platform, called SoftMealy, has been presented by Hsu and Dung [Hsu and Dung 1998]. SoftMealy was the rst wrapper induction system specically designed to work in the Web Data Extraction context. Relying on non-deterministic nite state automata (also known as nite-state transducers (FST)), SoftMealy uses a bottom-up inductive learning approach to learn extraction rules. During the training session the system acquires training pages represented as an automaton on all the possible permutations of Web pages: states represent extracted data, while state transitions represent extraction rules. SoftMealys main strength was its novel method of internal representation of the HTML documents. In detail, during a pre-processing step, each considered Web page was encoded into tokens (dened according to a set of inferential rules). Then, tokens were exploited to dene separators, considered as invisible borderlines between two consecutive tokens. Finally, the FST was fed by sequence of separators, instead of raw HTML strings (as in WIEN), so that to match tokens with contextual rules (dened to characterize a set of individual separators) to determine the state transitions. The advantages of SoftMealy with respect to WIEN are worth noting: in fact, the system was able to deal with a number of exception, such as missing values/attributes, multiple attribute values, variant attribute permutations and also with typos. STALKER. The last learning-based system discussed in this part is called STALKER [Muslea et al. 1999]. It was a supervised learning system for wrapper induction sharing some similarities with SoftMealy. The main dierence between these two systems is the specication of relevant data: in STALKER, a set of tokens is manually positioned on the Web page, so that to identify information that the user intend to extract. This aspect ensures the capability of STALKER of handling with empty values, hierarchical structures and non ordered items. This system models a Web page content by means of hierarchical relations, represented by using a tree data structure called embedded catalog tree (EC tree). The root of the EC tree is populated the sequence of all tokens (whereas, STALKER considers as token any piece of text or HTML tag in the document). Each child node is a sub-sequence of tokens inherited by its parent node. This implies that each parent node is a super-sequence of tokens of its children. The super-sequence is used, at each level of the hierarchy,
ACM Computing Surveys, Vol. V, No. N, July 2012.

20

Emilio Ferrara et al.

to keep track of the content in the sub-levels of the EC tree. The extraction of elements of interest for the user is achieved by inferring a set of extraction rules on the EC tree itself a typical example of extraction rule inferred by STALKER is the construct SkipTo(T), a directive that indicates, during the extraction phase, to skip all tokens until the rst occurrence of the token T is found. The inference of extraction rules exploits the concept of landmarks, sequences of consecutive tokens adopted to locate the beginning and the end of a given item to extract. STALKER is also able to dene wildcards, classes of generic tokens that are inclusive of more specic tokens. 2.3.1 Hybrid systems: learning-based wrapper generation. Wrapper generation systems discussed in Section 2.2 and wrapper induction techniques discussed in Section 2.3 dier essentially for two aspects: (1) The degree of automation of the Web Data Extraction systems; (2) The amount and the type of human engagement required for the functioning. The rst point is related to the ability of the system to work in an autonomous way, ensuring sucient standards of robustness and reliability, according to the requirements of users. Regarding the second point, most of the wrapper induction systems need labeled examples provided during the training sessions, thus requiring human expert engagement for the manual labeling phase. Wrapper generation systems, on the other hand, engage users into their maintenance, unless automatic techniques are employed, as those discussed in Section 2.2.2. Interestingly, a new class of platforms has been discussed in recent literature, that adopts a hybrid approach that sits between learning-based wrapper induction systems and wrapper generation platforms. The rst example of this class of systems is given by RoadRunner [Crescenzi et al. 2001; Crescenzi and Mecca 2004], a template-based system that automatically generates templates to extract data by matching features from dierent pages in the same domain. Another interesting approach is that of exploiting visual cues and spatial reasoning to identify elements in the Web pages with a Computer Vision oriented paradigm. This part concludes with the discussion of these two systems. Template-based matching. The rst example of hybrid system is provided by RoadRunner [Crescenzi et al. 2001; Crescenzi and Mecca 2004]. This system might be considered as an interesting example of automatic wrapper generator. The main strength of RoadRunner is that it is oriented to data-intensive websites based on templates or regular structures. The system tackles the problem of data extraction exploiting both features used by wrapper generators, and by wrapper induction systems. In particular, RoadRunner can work using information provided by users, in the form of labeled example pages, or also by automatically labeling Web pages (such as WIEN), to build a training set. In addition, it might exploits a priori knowledge on the schema of the Web pages, for example taking into account previously learned page templates. RoadRunner relies on the idea of working with two HTML pages at the same time in order to discover patterns analyzing similarities and dierences between structure and content of each pair of pages. Essentially, RoadRunner can extract relevant information from any Web site containing at least two Web pages with a
ACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

21

similar structure. Since usually Web pages are dynamically generated starting from template, and relevant data are positioned in the same (or in similar) areas of the page, RoadRunner is able to exploit this characteristic to identify relevant pieces of information, and, at the same time, taking into account small dierences due to missing values or other mismatches. The authors dened as class of pages those Web sources characterized by a common generation script. Then, the problem is reduced to extracting relevant data by generating wrappers for class of pages, starting from the inference of a common structure from the two-page-based comparison. This system can handle missing and optional values and also structural dierences, adapting very well to all kinds of data-intensive Web sources. Another strength of RoadRunner is its high-quality open-source implementation5 , that provides a high degree of reliability of the extraction system. Spatial reasoning. The paradigm of the Computer Vision has inspired also the eld of Web Data Extraction systems. In fact, a recent model of data extraction, called Visual Box Model, has been presented [Krpl et al. 2005; Gatterbauer and u Bohunsky 2006]. The Visual Box Model exploits visual cues to understand if, in the version of the Web page displayed on the screen, after the rendering of the Web browser, are present, for example, data in a tabular format. The advantage of this strategy is that it is possible to acquire data not necessarily represented by means of the standard HTML <table> format. The functioning of this technique is based on a X-Y cut OCR algorithm. This algorithm is able, given the rendered version of a Web page, of generating a visual grid, where elements of the page are allocated according to their coordinates determined by visual cues. Cuts are recursively applied to the bitmap image representing the rendering of the Web page, and stored into an X-Y tree. This tree is built so that ancestor nodes with leaves represent not-empty tables. Some additional operations check whether extracted tables contain useful information. This is done because although it is a deprecated practice many Web pages use tables for structural and graphical purposes, instead of for data representation scopes. The Visual Box Model data extraction system is implemented by means of an internal rendering engine that produces a visualization of the Web page relying on Gecko6 , the same rendering engine used by the Mozilla Web browser. By exploiting the CSS 2.0 box model, the algorithm is able to access the positional information of any given element. This is achieved by a bridge between the rendering engine and the application, implemented by means of the XPCOM library7 . 3. WEB DATA EXTRACTION SYSTEMS

In this section we get into details regarding the characteristics of existing Web Data Extraction systems. We can generically dene a Web Data Extraction system as a platform implementing a sequence of procedures (for example, Web wrappers) that extract infor5 http://www.dia.uniroma3.it/db/roadRunner/ 6 https://developer.mozilla.org/en/Gecko 7 https://developer.mozilla.org/it/XPCOM

ACM Computing Surveys, Vol. V, No. N, July 2012.

22

Emilio Ferrara et al.

mation from Web sources [Laender et al. 2002]. From this generic denition, we can infer two fundamental aspects of the problem: Interaction with Web pages. The rst phase of a generic Web Data Extraction system is the Web interaction [Wang et al. 2000]: the Web Data Extraction system accesses a Web source and extracts information stored in it. Web sources usually coincide with Web pages, but some approaches consider also as RSS/Atom feeds [Hammersley 2005] and Microformats [Khare and Celik 2006]. Some commercial systems, Lixto for rst but also Kapow Mashup Server (described below), include a Graphical User Interface for fully visual and interactive navigation of HTML pages, integrated with data extraction tools. The most advanced Web Data Extraction systems support the extraction of data from pages reached by deep Web navigation [Baumgartner et al. 2005]: they simulate the activity of users clicking on DOM elements of pages, through macros or, more simply, by lling HTML forms. These systems also support the extraction of information from dynamically generated Web pages, usually built at run-time as a consequence of the user request, lling a template page with data from some database. The other kind of pages are commonly called static Web pages, because of their static content. OxPath [Furche et al. 2011], which is part of the DIADEM project [Furche et al. 2012] is a declarative formalism that extends XPath to support deep Web navigation and data extraction from interactive Web sites. It adds new location steps to simulate user actions, a new axis to select dynamically computed CSS attributes, a convenient eld function to identify visible elds only, page iteration with a Kleene-star extension for repeated navigations, and new predicates for marking expressions for extraction identication. Generation of a wrapper. A Web Data Extraction system must implement the support for wrapper generation and wrapper execution. Another denition of Web Data Extraction system was provided by Baumgartner et al. [Baumgartner et al. 2009]. They dened a Web Data Extraction system as a software extracting, automatically and repeatedly, data from Web pages with changing contents, and that delivers extracted data to a database or some other application. This is the denition that better ts the modern view of the problem of the Web Data Extraction as it introduces three important aspects: Automation and scheduling Data transformation, and the Use of the extracted data In the following we shall discuss each of these aspects into detail. Automation and Extraction. Automating the access to Web pages as well as the localization of their elements is one of the most important features included in last Web Data Extraction systems [Phan et al. 2005]. The capability to create macros to execute multiple instances of the same task, including the possibility to simulate the click stream of the user, lling forms and selecting menus and buttons, the support for AJAX technology [Garrett 2005] to handle the asynchronous updating of the page, etc. are only some of the most important automation features.
ACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

23

Also the scheduling is important, for example if a user wants to extract data from a news Web site updated every 5 minutes, many of the last tools let him to setup a scheduler, working like a cron, launching macros and executing scripts automatically and periodically. Data transformation. Information could be wrapped from multiple sources, which means using dierent wrappers and also, probably, obtaining dierent structures of extracted data. The steps between extraction and delivering are called data transformation: during these phases, such as data cleaning [Rahm and Do 2000] and conict resolution [Monge 2000], users reach the target to obtain homogeneous information under a unique resulting structure. The most powerful Web Data Extraction systems provide tools to perform automatic schema matching from multiple wrappers [Rahm and Bernstein 2001], then packaging data into a desired format (e.g., a database, XML, etc.) to make it possible to query data, normalize structure and de-duplicate tuples. Use of extracted data. When the extraction task is complete, and acquired data are packaged in the needed format, this information is ready to be used; the last step is to deliver the package, now represented by structured data, to a managing system (e.g., a native XML DBMS, a RDBMS, a data warehouse, a CMS, etc.). In addition to all the specic elds of application covered later in this work, acquired data can be also generically used for analytical or statistical purposes [Berthold and Hand 1999] or simply to republish them under a structured format. 3.1 Layer cake comparisons

In this section, we summarize the capability stacks of Web data extraction systems from our understanding, including aspects of wrapper generation, data extraction capabilities, and wrapper usage. In particular, we introduce some specic aspects and illustrate the technological evolution of Web Data Extraction systems according to each aspect. We use ad hoc diagrams structured as a group of layers (layer cakes). In these diagrams, bottom (resp., top) layers correspond to the earliest (resp., latest) technological solutions. Wrapper Generation: Ease of Use (Figure 3). The rst approaches to consuming facts from the Web were implemented by means of general purpose languages. Over time libraries (e.g. Ruby Mechanize) and special-purpose query languages evolved on top of this principle (e.g., Jedi [Huck et al. 1998] and Florid [May et al. 1999]). Wizards that simplify the way to specify queries are the next logical level and for instance have been used in W4F [Sahuguet and Azavant 1999] and XWrap [Liu et al. 2000]. Advanced Web Data Extraction systems oer GUIs for conguration, either client-based (e.g. Lapis), Web-based (e.g. Dapper and Needlebase) or as browser extensions (e.g. iOpus and Chickenfoot). Commercial frameworks oer a full IDE with functionalities described in the previous section (e.g. Denodo, Kapowtech, Lixto and Mozenda). Wrapper Generation: Creation Paradigma (Figure 4). From the perspective of how the system supports the wrapper designer to create robust extraction programs, the simplest approach is to manually specify queries and test them against sample sites individually. Advanced editors oer highlighting of query keywords
ACM Computing Surveys, Vol. V, No. N, July 2012.

24

Emilio Ferrara et al.

Fig. 3.

Layer Cake: Wrapper Generation: Ease of Use

and operators and assist query writing with auto-completion and similar usability enhancements (e.g., Screen-Scraper). In case of procedural languages, debugging means and visual assistance for constructs such as loops are further means to guided wrapper generation. Regarding the denition of deep Web navigations, such as form llouts, a number of tools oer VCR-style recording the human browsing and replaying the recorded navigation sequence (e.g., Chickenfoot, iOpus, Lixto). Visual and interactive facilities are oered by systems to simplify fact extraction. Users mark an example instance, and the system identies the selected element in a robust way, and possibly generalize to match further similar elements (e.g., to nd all book titles). Such means are often equipped with Machine Learning methods, in which the user can select a multitude of positive and negative examples, and the system generates a grammar to identify the Web objects under consideration, often in an iterative fashion (e.g. Wien, Dapper, Needlebase). Click and drag and drop features further simplify the interactive and visual mechanisms. Finally, some systems oer vertical templates to easily create wrappers for particular domains, for example for extracting hotel data or news item, using Natural Language Processing techniques and domain knowledge, or for extracting data from typical Web layouts, such as table structures or overview pages with next links. Deep Web Navigation Capabilities (Figure 5). Before the advent of Web 2.0 techniques, dynamic HTML and AJAX it was usually sucient to consider the Web as a collection of linked pages. In such cases, form lling can be simulated by tracking the requests and responses from the Web Server, and replaying the sequence of requests (sometimes populating a session id dynamically, extracted from a previous page of the sequence). Alternatively, early Web Data extraction systems have been inuenced by screen scraping technologies like they were being used for automating 3270 applications or like used for automating native applications, usually relying heavily on coordinates. Understanding and replaying DOM Events on Web objects is the next logical level in this capability stack. Advanced systems even go a step further, especially when embedding a full browser: the click on an element is recorded in a robust way and during replay the browser
ACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

25

Vertical Templates

Click and Drag Machine Learning

Form Iteration Screen Scraping Parametrization Query Probing

Visual Macro Recording and Interactive Visual Generation

Visual View on Procedural Program

Fig. 4.

Layer Cake: Wrapper Generation: Creation Paradigma

is informed to do a visual click on such an element, handing the DOM handling over to the browser and making sure that the Web page is consumed exactly in the way the human user consumes it. Orthogonal to such features are capabilities to parametrize deep Web sequences, and to use query probing techniques to automate deep Web navigation to unknown forms.

Human-like Macros Click Recording

DOM Events Request/Response Sequence of Links

Fig. 5.

Layer Cake: Deep Web Navigation Capabilities

Web Data Extraction Capabilities (Figure 6). Over time, various approaches to modeling a Web page have been discussed. The simplest way is the to work on the stream received from the Web server, for example using regular expressions. In some cases, this is sucient and even the preferred approach in large-scale scenarios due to avoiding to build up a complex and unperformant model. On the other hand, in complex Web 2.0 pages or in pages that are not well-formed, it can be extremely cumbersome to work on the textual level only. Moreover, such wrappers are not very maintainable and break frequently. The most common approach is to work on the DOM tree or some other kind of tree structure. This approach has been followed both by the academic community and by commercial approaches. In the academic communities studying expressiveness of language
ACM Computing Surveys, Vol. V, No. N, July 2012.

26

Emilio Ferrara et al.

over trees and tree automata are interesting, and from a commercial point of view, it is convenient and robust to use languages such as XPath for identifying Web objects. Usually, not only the elements of a DOM tree are considered, but also the events, enabling to specify data extraction and navigation steps with the same approach. In the Web of today, however, very often the DOM tree does not really capture the essential structure of a Web page as presented to the human user in a browser. A human perceives something as table structure, whereas the DOM tree contains a list of div elements with absolute positioning. Moreover, binary objects embedded into Web pages such as Flash pose further challenges and are not covered with a tree structure approach. Hence, screen-scraping made back its way into novel Web Data extraction frameworks, using methods from document understanding and spatial reasoning such as the approaches of the TamCrow project [Krpl-Sypien et al. 2011], of the ABBA project [Fayzrakhmanov et al. u 2010] spatial XPath extensions [Oro et al. 2010] and rendition-based extensions in RoadRunner to detect labels [Crescenzi et al. 2004].

Application Logic Workflow Visual Rendition HTML Tree Plain Textual Strings
Fig. 6. Layer Cake: Web Data Extraction Capabilities

Interactive DOM Tree

Parser and Browser Embedding (Figure 7). This capability stack is closely related to the Deep Web capabilities but focussing on the technical realization of parsing and browser embedding. Simple approaches create their own parser to identify relevant HTML tags, more sophisticated approaches use DOM libraries without an associated browser view. Due to the fact that many Web Data Extraction frameworks are implemented in Java, special-purpose browsers such as the Java Swing browser and the ICE browser are and have been used. The most powerful approaches are the ones that embed a standard browser such as IE, Firefox or WebKit-based browsers. In case of Java implementations, interfaces such as Java-XPCOM bridge or libraries such as JRex are used to embed the Mozilla browser. Embedding a full browser not only gives access to the DOM model, but additionally to other models useful for data extraction, including the CSS Box model. Some tools go a dierent direction and instead of embedding a browser, they are implemented as browser extension, imposing some restrictions and inconveniences. Orthogonal to the browser stack are the capabilities to
ACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

27

extend extraction functionalities to unstructured text parsing exploiting Natural Language Processing techniques.

DOM Parser Simple Text/Tag Parser

Fig. 7. Layer Cake: Parser and Browser Embedding

Complexity of Supported Operations (Figure 8). Simple data extraction tools offer Web notication, e.g.if a particular word is mentioned. Web macro recorders allow users to create deep bookmarks, and Web clipping frameworks clip fragments of a Web page to the desktop. Personalization of a Web page (for instance, some tools oer to alter the CSS styles to make frequently used links more prominent on a page) is a further level in the layer cake. Batch processing frameworks oer functionalities to replay ample extraction tasks (e.g., running through many dierent values in form llout). Advanced systems go a step further and use sophisticated extraction plans and anonymization techniques that ensure to stay under the radar and not harm target portals with too many requests at once, and deliver the data into further applications such as market intelligence platforms. The expressiveness of a wrapper language contributes as well to the complexity of supported operations.

Batch Processing End User Personal Views Web Macro Recorder Web Clipping

Web Notification Tools

Fig. 8. Layer Cake: Complexity of Supported Operations

ACM Computing Surveys, Vol. V, No. N, July 2012.

Expressiveness of Wrapper Language

Enterprise Process Integration

NLP / Text Understanding

Standard Browser (IE, FF, etc.) Special-Purpose Browser (Swing, ICE, etc.)

28

Emilio Ferrara et al.

4.

APPLICATIONS

In the literature of the Web Data Extraction discipline, many works cover approaches and techniques adopted to solve some particular problems related to a single or, sometimes, a couple of elds of application. The aim of the second part of this paper is to survey and analyze a large number of applications that are strictly interconnected with Web Data Extraction tasks. To the best of our knowledge, this is the rst attempt to classify applications based on Web Data Extraction techniques even if they have been originally designed to operate in specic domain and, in some cases, they can appear as unrelated. The spectrum of applications possibly beneting from Web Data Extraction techniques is quite large and it encompasses applications designed to work in the business domain to applications developed in the context of the Social Web. We classied applications of Web Data Extraction techniques in two main categories: Enterprise Applications: applications falling in this category are mainly conceived with a commercial goal and they often aim at increasing both the level of automation of a business process and its eciency. Social Web Applications: applications falling in this category are mainly designed to extract and collect data from a Social Web platform (e.g., a Social Networking Website like Facebook or a resource sharing systems like Flickr). The classication scheme above, however, has not to be intended as hard, i.e., there are some applications which have been designed to work in the enterprise context but they share some features with applications working on Social Web platforms. For instance, several rms provide through their customer care services tools like discussion forums allowing a customer to directly communicate with the rm or to get in touch and share opinions/comments with other customers. Messages exchanged in a discussion forum can be extracted and analyzed to identify what are the trending discussion topics as well as to measure the level of satisfaction perceived by the customers. This information is valuable because it enables rm managers to design strategies to increase the quality of services/products delivered to customers. Of course, applications working on data produced in discussion forums have also a strong social dimension because they deal with data generated by users and some of this information derive often from social interactions (think of the volume of information generated when a user poses a question to the community and other forum members answer her questions). In addition to classifying applications into the enterprise and Social Web categories, we can introduce other criteria which rely on the fact that data are extracted from a single or multiple sources, data can be of the same format or, nally, are associated with the same domain. In particular, we introduce the following criteria: Single Source vs. Multiple Sources. Web Data Extraction techniques can be applied on data residing on a single platform or, vice versa, they can collect data located in dierent platforms. As an example, in the context of Enterprise applications, some applications fetch data from a single platform: a relevant example is provided by applications to manage customer care activities. As previously pointed out, these application
ACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

29

crawls a corpus of text documents produced within a single platform. By contrast, in some cases, the application can benet of data scattered across multiple system. In the context of Enterprise applications, a notable example of application gathering data from multiple sources is given by Comparison Shopping: in such a case, a Web Data Extraction technique is able to collect data associated with a target product from a multitude of e-commerce platforms and to compare, for instance, the price and the shipping details of that product in each of these platform. The classication above is particularly interesting in the eld of Social Web applications: in fact we discuss separately applications designed to crawl and collect data from a single platform (see Section 4.2.1) from applications running on multiple Social Web platforms. Homogeneity in Format vs. Heterogeneity in Format. The second classication criterium we introduce answers the following question: does the application collect data of the same format or, vice versa, can it collect data of dierent formats? On the basis of such a criterium we can classify existing applications in homogeneous-format and heterogeneous-format. As for enterprise applications, a nice example of homogeneous-format applications is given by Opinion Sharing Applications. This term identies the applications devoted to collect the opinions a user expresses about a given service/product. These opinions are typically represented by short textual reviews or blog post. In such a case, the type of data extracted by a Web Data Extraction technique is a string. In some cases, users are allowed to provide scores and the format of extracted data is a discrete value, generally ranging in an interval like [0,5]. In case of Social Web applications, many approaches (like [Gjoka et al. 2010; Catanese et al. 2011; Mislove et al. 2007; Ye et al. 2010]) are devoted to extract friendship relationships from a Social Networking Websites. In such a case, extracted data are of the same format that can be interpreted as a triplet of the form ux , uy , f being ux and uy the user identiers of two Social Network members; the value f can be a boolean (e.g., it can be true if ux and uy are friends, false otherwise) or it can be a numerical value (e.g., it can be set equal to the number of messages ux sent uy ). Some applications are designed to collect data of dierent type. A relevant example in the eld of enterprise applications is provided by those applications designed to manage Business Intelligence tasks. These application are able to collect data of dierent type like (numerical data or textual ones) as well as to manage both structured data (e.g., data coming from relational tables of a DBMS) as well as unstructured ones (e.g., text snippets present in HTML pages). Relevant examples of applications capable of extracting data of dierent type are also present in the context of Social Web applications: for instance, [Kwak et al. 2010] performed in 2009 a crawl of the whole Twitter platform which produced textual data (e.g., the tweets and re-tweets produced by users) as well as data indicating the dierent types of connections among users (following, reply to and mention) [Romero and Kleinberg 2010]. Single Purpose or Multiple Purpose. In some cases, the goal of a Web Data Extraction tool is to extract data describing a specic facet of a particular the
ACM Computing Surveys, Vol. V, No. N, July 2012.

30

Emilio Ferrara et al.

same social phenomenon or business process. In such a case, an application is catalogued as unique purpose. Other applications, instead, aim at collecting data of dierent nature which, if properly linked, are relevant to better understand and interpret a particular phenomenon. After linking data, novel and sophisticated applications running on them can be dened. Applications belonging to this category will be called multi purpose. As an example of single purpose applications, we focus on applications devoted to collect bibliographic data as well as citations among papers. To collect bibliographical information, Web Data Extraction techniques are required to query several databases containing scientic publications (like ISI Web of Science, SCOPUS, PubMed and so on) but the goal of the application, however, is to collect all the citations associated with a particular paper or to a particular author because this information is useful to assess the impact of an author or a scientic publication in a scientic discipline. In the eld of Social Web applications, an example of single purpose application is given by applications devoted at collecting data about human activities in dierent Social Web platforms: for instance, the tags contributed by a user in dierent systems [Szomszor et al. 2008]. This is relevant to understand if the language of a user is uniform across various platforms or if the platform features impact on the user formation of the vocabulary. We can cite some relevant examples of multi purpose applications both at the enterprise and at the Social Web level. For instance, in the class of enterprise applications, some applications are able to collect data produced by dierent Web Services and combine these data to produce more advanced applications. For instance, in the travel industry, one can think of applications collecting data on ights and hotels and combine these data to produce holiday packages. In the eld of Social Web applications, several authors observed that the merging of dierent type of data describing dierent type of human activities produces a more detailed knowledge of human needs and preferences. For instance, in [Schifanella et al. 2010] the authors conducted extensive experiments on data samples extracted from Last.Fm and Flickr. They showed that a a strong correlations exist between user social activities (e.g., the number of friends of a user or the number of groups she joined) and the tagging activity of the same user. A nice result provided in [Schifanella et al. 2010] was that user contributed tags are a useful indicator to predict friendship relationships. Other studies, involving the geographical location of a user with the content she posted (e.g., photos) are provided in [Crandall et al. 2009] and [Kinsella et al. 2011]. 4.1 Enterprise Applications

In thsi section we describe the main features of software applications and procedures related with Web Data Extraction with a direct, subsequent or nal commercial scope. 4.1.1 Context-aware advertising. Context-aware advertising techniques aim at presenting to the end user of a Web site commercial thematized advertisements together with the content of the Web page the user is reading. The ultimate goal is to increase the value that a Web page can have for its visitors and, ultimately, to raise the level of interest in the ad.
ACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

31

First eorts to implement context-aware advertising were made by Applied Semantic, Inc.8 ; subsequently Google bought their AdSense advertising solution. The implementation of context-aware advertisements requires to analyze the semantic content of the page, extract relevant information, both in the structure and in the data, and then contextualize the ads content and placement in the same page. Contextual advertising, compared to the old concept of Web advertising, represents a intelligent approach to providing useful information to the user, statistically more interested in thematized ads, and a better source of income for advertisers. 4.1.2 Customer care. Usually medium- and big-sized company, with customers support, handle a large amount of unstructured information available as text documents. Relevant examples are emails, support forum discussions, documentation, shipment address information, credit card transfer reports, phone conversation transcripts, an so on. The ability of analyzing these documents and extracting the main concepts associated with them provides several concrete advantages. First of all, documents can be classied in a more eective fashion and this makes their retrieval easier. In addition, once the concepts present in a collection of documents have been extracted, it is possible to identify relevant associations between documents on the basis of the concepts they share. Ultimately, this enables to perform sophisticated data analysis targeted at discovering trends or, in case of forms, hidden associations among the products/services oered by a brand. In this scenario, Web Data Extraction techniques play a key role because they are required to quickly process large collections of textual documents and derive the information located in these documents. The retrieved data are nally processed by means of algorithms generally coming from the area of Natural Language Processing. 4.1.3 Database building. In the Web marketing sector, Web Data Extraction techniques can be employed to gather data referring to a given domain. These data may have a twofold eect: (i) a design, through reverse engineering analysis, can design and implement a DBMS representing that data; (ii) the DBMS can be automatically populated by using data provided by the Web Data Extraction system. The activity of designing and building a DBMS starting from collected data will be called Database Building. Fields of application are countless: nancial companies could be interested in extracting nancial data from the Web and storing them in their DBMSs. Extraction tasks are often scheduled in such a way as to be executed automatically and periodically. Also the real estate market is very orid: acquiring data from multiple Web sources is an important task for a real estate company, for comparison, pricing, co-oering, etc. Companies selling products or services probably want to compare their pricing with other competitors: products pricing data extraction is an interesting application of Web Data Extraction systems. Finally we can list other related tasks, obviously involved in the Web Data Extraction: duplicating an on-line database,
8 http://www.appliedsemantics.com/

ACM Computing Surveys, Vol. V, No. N, July 2012.

32

Emilio Ferrara et al.

extracting dating sites information, capturing auction information and prices from on-line auction sites, acquiring job postings from job sites, comparing betting information and prices, etc. 4.1.4 Software Engineering. Extracting data from Web sites became interesting also for Software Engineering: for instance, Rich Internet Applications (RIAs) are rapidly emerging as one of the most innovative and advanced kind of application on the Web. RIAs are Web applications featuring a high degree of interaction and usability, inherited from the similarity to desktop applications. Amaltano et al. [Amaltano et al. 2008] have developed a reverse engineering approach to abstract Finite States Machines representing the client-side behavior oered by RIAs. 4.1.5 Business Intelligence and Competitive Intelligence. Baumgartner et al. [Baumgartner et al. 2005; Baumgartner et al. 2009; Baumgartner et al. 2010] deeply analyzed how to apply Web Data Extraction techniques and tools to improve the process of acquiring market information. A solid layer of knowledge is fundamental to optimize the decision-making activities and an large amount of public information could be retrieved on the Web. They illustrate how to acquire these unstructured and semi-structured information. In particular, using the Lixto Suite to access, extract, clean and deliver data, it is possible to gather, transform and obtain information useful to business purposes. It is also possible to integrate these data with other common platforms for Business Intelligence, like SAP 9 or Microsoft Analysis Services [Melomed et al. 2006]. Wider, the process of gathering and analyzing information about products, customers, competitors with the goal of helping the managers of a rm in decisional processes is commonly called Competitive Intelligence, and is strictly related to data mining [Han and Kamber 2000]. Zanasi [Zanasi 1998] was the rst to introduce the possibility of acquiring these data, through data mining processes, on public domain information. Chen et al. [Chen et al. 2002] developed a platform, that works more like a spider than a Web Data Extraction system, which represents a useful tool to support Competitive Intelligence operations. In Business Intelligence scenarios, we ask Web Data Extraction techniques to satisfy two main requirements: scalability and ecient planning strategies because we need to extract as much data as possible with the smallest amount of resources in time and space. 4.1.6 Web process integration and channel management. In the Web of today data is often available via APIs (e.g., refer to Programmable Web 10 ). Nevertheless, the larger amount of data is primarily available in semi-structured formats such as HTML. To use Web data in Enterprise Applications and service-oriented architectures, it is essential to provide means for automatically turning Web Applications and Web sites into Web Services, allowing a structured and unied access to heterogeneous sources. This includes to understand the logic of the Web application, to ll out form values, and to grab relevant data. In a number of business areas, Web applications are predominant among business
9 http://www.sap.com 10 http://www.programmableweb.com/

ACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

33

partners for communication and business processes. Various types of processes are carried out on Web portals, covering activities such as purchase, sales, or quality management, by manually interacting with Web sites. Typical vertical examples, where Web Data Extraction proves useful include channel management in the travel industry (like automating the regular oerings of rooms on hotel portals with bidirectional Web connectors), re-packaging complex Web transactions to Web services and consequently to other devices, as well as automating communication of automotive suppliers with automotive companies. Tools for wrapper generation pave the way for Web Process Integration and enable the Web of Services, i.e., the seamless integration of Web applications into a corporate infrastructure or service oriented landscape by generating Web services from given Web sites [Baumgartner et al. 2010]. Web process integration can be understood as front-end and outside-in integration: integrate cooperative and non-cooperative sources without the need for information provider to change their backend. Additional requirements in such scenarios include to support a large number of users and real-time parametrized Web queries and support of complex Web transactions. 4.1.7 Functional Web application testing. Testing and Quality Management are essential parts of the life-cycle of software. Facets of testing are manifold, including functional tests, stress/load tests, integration tests, and testing against specications, to name a few. Usually, the strategy is to automate a large percentage of functional tests and execute test runs as part of nightly builds as regression tests. Such tests occur at various levels, for example testing the system functionality as a blackbox via APIs, or testing the system at the GUI level simulating either the users steps or creating a model of possible application states. In todays world of Software-as-a-Service platforms and Web oriented architectures, Web application testing plays an important role. One aspect is simulating the users path through the application logic. Robust identication criteria can be created by taking advantage of the tree structure or visual structure of the page. Typical actions in such test scripts include to set/get values of form elds, picking dates, checkpoints to compare values, and following dierent branches depending on the given page. Due to automation, every step can be parametrized and a test script executed in variations. The requirements for tools in the area of Web application testing are to deal well with AJAX/dynamic HTML, to create robust test scripts, to eciently maintain test scripts, to execute test runs and create meaningful reports, and, unlike other application areas, the support of multiple state-of-the-art browsers in various versions is an absolute must. One widely used open source tool for Web application testing is Selenium 11 . 4.1.8 Comparison shopping. One of the most appreciated services in e-commerce area is the comparison shopping, i.e., the capability of comparing products or services; various type of comparisons are allowed going from simple prices comparison to features comparison, technical sheets comparison, user experiences comparison, etc.
11 http://seleniumhq.org/

ACM Computing Surveys, Vol. V, No. N, July 2012.

34

Emilio Ferrara et al.

These services heavily rely on Web Data Extraction, using Web sites as sources for data mining and a custom internal engine to make possible the comparison of similar items. Many Web stores today also oer personalization forms that make the extraction tasks more dicult: for this reason many last-generation commercial Web Data Extraction systems (e.g., Lixto, Kapow Mashup Server, UnitMiner12 , Bget13 ) provide support for deep navigation and dynamic content pages. 4.1.9 Mashup scenarios. Today, leading software vendors provide mashup platforms (such as Yahoo! Pipes or Lotus Mashups) and establish mashup communication standards such as EMML14 . A mashup is a Web site or a Web application that combines a number of Web sites into an integrated view. Usually, the content is taken via APIs, embedding RSS or Atom Feeds in a REST-like way. With wrapper technology, one can leverage legacy Web applications to light-weight APIs such as REST that can be integrated in mashups in the same fashion. Web Mashup Solutions no longer need to rely on APIs oered by the providers of sites, but can extend the scope to the whole Web. In particular, the deep Web gets accessible by encapsulating complex form queries and application logic steps into the methods of a Web Service. End users are put in charge of creating their own views of the Web and embed data into other applications (consumers as producers), usually in a light-weight way. This results in situational applications, possibly unreliable and unsecure applications, that however help to solve an urgent problem immediately. In Mashup scenarios, one important requirement of Web Data Extraction tools is the ease of use for non-technical content managers, to give them the possibility to create new Web connectors without help of IT experts. 4.1.10 Opinion mining. Related to comparison shopping, the opinion sharing represents its evolution: users want to express opinions on products, experiences, services they enjoyed, etc. The most common form of opinion sharing is represented by blogs, containing articles, reviews, comments, tags, polls, charts, etc. All this information usually lacks of structure, so their extraction is a big problem, also for current systems, because of the billions of Web sources now available. Sometimes model-based tools t good, taking advantage of common templates (e.g., Wordpress15 , Blogger16 , etc.), other times Natural Language Processing techniques t better. Kushal et al. [Dave et al. 2003] approached the problem of opinion extraction and subsequent semantic classication of reviews of products. Another form of opinion sharing in semi-structured platforms is represented by Web portals that let users to write unmoderated opinions on various topics.

12 http://www.qualityunit.com/unitminer/ 13 http://www.bget.com/ 14 http://www.openmashup.org 15 http://wordpress.org/ 16 https://www.blogger.com/

ACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

35

4.1.11 Citation databases. Citation databases building is an intensive Web Data Extraction eld of application: CiteSeer17 , Google Scholar18 and DBLP19 , amongst others, are brilliant examples of applying Web Data Extraction to approach and solve the problem of collect digital publications, extract relevant data, i.e., references and citations, and build structured databases, where users can perform searches, comparisons, count of citations, cross-references, etc. Several challenges are related to this context of application: for example, the corpus of scientic publications could rapidly vary over time and, to keep the database updated it could be necessary to repeatedly apply a Web Data Extraction process from scratch on the same Web sources. Such an operation, however, can be excessively time-consuming. An attempt to address this challenge has been done in [Chen et al. 2008]: in that paper, the authors suggest an incremental solution which requires to identify portions of information shared by consecutive snapshots and to reuse the information extracted from a snapshot to the subsequent one. 4.1.12 Web accessibility. Techniques for automatic data extraction and document understanding are extremely helpful in making Web pages more accessible to blind and partial-sighted users. Todays solution approaches are inecient to overcome the problem. The rst approach, screen-reader usage, is optimized for native client interfaces, and not well equipped to deal with the presentation, content and interactions in Web 2.0 such as understanding the reading order, telling the user what a date picker means, or jumping from one forum post to the next. The second approach, the Web Accessibility Initiative, is no doubt absolutely necessary and has dened valuable concepts, such as ARIA roles assignable to GUI elements. However, due to the additional investment at best such guidelines are applied in governmental sites. Approaches such as ABBA [Fayzrakhmanov et al. 2010] overcome these limitations. In the ABBA approach, a Web page is transformed into a formal multi-axial semantic model; the dierent axes oer means to reason on and serialize the document by topological, layout, functional, content, genre and saliency properties. A blind person can navigate along and jump between these axes to skip to the relevant parts of a page. E.g., the presentational axis contains transformed visual cues, allowing the user to list information in the order of visual saliency. 4.1.13 Main content extraction. Typical Web pages, for example news articles, contain, additionally to the main content, navigation menus, advertisements and templates. In some cases, such as when archiving a news article Web page for later oine reading it is convenient to get rid of such irrelevant fragments. To extract the main content only, one needs to apply techniques to distinguish the relevant content from the irrelevant one. Approaches range from complex visual Web page analysis to approaches leaning on text density or link density analysis. An approach for boilerplate detection using shallow text features was introduced in [Kohlschtter et al. 2010]. Additionally, tools/apps such as InstaPaper or the u
17 http://citeseer.ist.psu.edu/ 18 http://scholar.google.com/ 19 http://www.informatik.uni-trier.de/ley/db

ACM Computing Surveys, Vol. V, No. N, July 2012.

36

Emilio Ferrara et al.

Readability Library20 use main content extraction to store the relevant fragment and text from a Web page that resembles the article for later reading. 4.1.14 Web (experience) archiving. Digital preservation and Web data curation are the goals of the discipline of Web Archiving. On the one hand, this means to access information no longer available in the live Web, and on the other hand also to reect how the Web was used in former times. Events such as iPres and IWAW, and consortia such as Netpreserve, as well as various local Web archiving initiatives, tackle this task. There are numerous challenges [Masan`s 2006] due to the facts e that Web pages are ephemeral and due to unpredictable additions, deletions and modications. Moreover, the hidden Web poses a further challenge. Approaches to Web Archiving are manifold, and range from Web Crawling/Harvesting, server-side archiving, transaction-based archiving, archiving the content of Web databases to library-like approaches advocating persistent identiers (e.g., Digital Object Identier - DOI). Especially in the use case about Web database content archiving, Web data extraction techniques are exploited. Another possibility of archiving the Web is to archive how Web pages are consumed. We can consider this as a type of event-based archiving, that makes sense especially for rich Web applications. The idea is not to archive everything, but to archive selectively sample pathes through an application. This requires to choose sample sites, get an understanding about common and frequent pathes through an application and store the interaction sequence. In a museum-like approach, such selected sequences are stored and can be restored or replayed to provide the user with the experience how the Web was consumed. Deep Web navigation techniques and form understanding are the key technologies here. 4.1.15 Summary. Figure 9 summarizes the 14 discussed enterprise application scenarios. For each scenario we describe the main ingredients of the value chain, as illustrated in the scenario descriptions. 4.2 Social Web Applications

In the latest years, Social Web platforms emerged as one of the most relevant phenomenon on the Web: these platforms are built around users, letting them to create a web of links between people, to share thoughts, opinions, photos, travel tips, etc. In such a scenario, often called Web 2.0 users turn from passive consumers of contents to active producers. Social Web platforms provide novel and unprecedented research opportunities. In fact, The analysis on a large, often planetary, scale of patterns of users interactions provides the concrete opportunity of answering questions like these: how does human relationships (e.g., friendship relationships) are created and evolve over time [Kleinberg 2008]? How does novel ideas spread and propagate through the web of human contacts [Bettencourt et al. 2006]? How does the human language evolve through social interactions (e.g., how do person expand their lexicon on the basis of their interactions with other persons) [Mathes 2004]? Besides scientic questions, the analysis of patterns of human interactions in Social Web platforms has also relevant implication at the business level: if we are
20 The

reader is referred to the PHP port http://www.keyvan.net/2010/08/php-readability/

ACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

37

Context-Aware Advertisting Customer Care Database Building Software Engineering Competitive Intelligence Web Process Integration Web Application Testing Comparison Shopping Mashups Opinion Mining Citation Databases Web Accessibility Main Content Extraction Web Experience Archiving

NER/NLP Search Search Libraries Deep Web Processes Deep Web Processes Deep Web Processes Deep Web Processes Fact Extraction Search Fact Extraction Fact Extraction Text Extraction Search

Classification NER/NLP Fact Extraction Classification Schema Matching

Fact Extraction

Data Cleansing and Integration

Record Linkage

Data Mining

Reports/ Analytics

Fact Bi-directional Extraction Enterprise integration Comparisons and Assertions Fact Extraction Light-weight APIs NER/NLP Data Integration Content Repackaging Data Mining Test Management Process Repackaging

Process Repackaging

Classification

Deep Web Processes

Experience Archiving

Fig. 9.

Summary of the application domains in the enterprise context.

able to understand the dynamics of interactions among humans, we are also able to identify how groups of users aggregate themselves around shared interests. This is a crucial step for marketing purposes: once users have been grouped, we can, for instance, selectively disseminate commercial advertisements only to those groups formed by users who are actually interested in receiving those advertisements. In an analogous fashion, the fabric of social interactions can be used to identify inuential users, i.e., those users whose commercial behaviors are able to stimulate the adoption/rejection of a given product by large masses of users. Finally, Social Web users often create accounts and/or proles in multiple platforms [Szomszor et al. 2008; De Meo et al. 2009]. Correlating these accounts and proles is a key step to understand how the design features and the architecture of a Social Web platform impact on the behavior of a user: so, for instance, one may ask whether some functionalities provided by a given platform augment the aptitude of users to socialize or they impact on the volume of contents produced by a user. Once the relationship between the features of a given platform and the behavior of a user has been elucidated, the designers/managers of that platform can provide novel services to raise the level of engagement of a user in the platform or to raise their degree of loyalty (e.g., to avoid users become inactive in the platform and migrate to other ones).
ACM Computing Surveys, Vol. V, No. N, July 2012.

38

Emilio Ferrara et al.

In the context above, Web Data Extraction techniques play a key role because the capability of timely gathering large amounts of data from one or more Social Web platforms is a an indefeasible tool to analyze human activities. Traditional Web Data Extraction techniques are challenged by new and hard problems both at the technical and scientic level. First of all, Social Web platforms feature a high level of dynamism and variability because the web of contacts of a user or a group of users may signicantly vary in small time slots; therefore, we need to design Web Data extraction algorithms/procedures capable of gathering large amount of data in a quick fashion, in such a way as to the fragments of collected data keep pace with changes occurring in the structure of the user social network. If such a requirement is not satised, the picture emerging from the analysis we can carry out on the data at our disposal could be wrong and it would fail to capture the structure and evolution of human interactions in the platform. A second challenge depends on the fact that Web Data Extraction algorithms are able to capture only a portion of the data generated within one or more platforms. Therefore, we are required to check that the features of a data sample generated as the output of a Web Data Extraction algorithm replicate fairly well the structure of the original data within the platform(s). Finally, since data to gather are associated with humans or reect human activities, Web Data Extraction techniques are challenged to provide solid guarantees that user privacy is not violated. In the remaining of this section we will provide an answer to these questions and illustrate how Web Data Extraction techniques have been applied to collect data from Social Web platforms. First of all, we describe how to gather data about user relationships and activities within a single social platform (see Section 4.2.1). Secondly, we consider users who created and maintain an account in multiple Social Web platforms and discuss issues and challenges related to data collection in this scenario (see Section 4.2.2). 4.2.1 Extracting data from a single Online Social Web platform. In this section we rst discuss the technological challenges arising if we aim at collecting data about social relationships, user activities and the resources produced and shared by users. Subsequently, we discuss privacy risks associated with the consumption/usage of human related data. Technical challenges for collecting data from a single Social Web platform. We can classify techniques to collect data from a Social Web platform into two main categories: the former category relies on the usage of ad-hoc APIs, usually provided by the Social Web platform itself; the latter relies on HTML scraping. As for the rst category of approaches, we point out that, today, Social Web platforms provide powerful APIs (often available in multiple programming languages) allowing to retrieve in an easy and quick fashion a wide range of information from the platform itself. This information, in particular, regards not only social connections involving members of the platforms but also the content the users posted and, for instance, the tags they applied to label available content. We can cite the approach of [Kwak et al. 2010] as a relevant example of how to collect data from a Social Web platform by means of an API. In that paper, the authors present the results of the crawling of the whole Twitter platform. The
ACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

39

dataset described in [Kwak et al. 2010] built consisted of 41.7 million user proles, 1.47 billion social relations; in addition to collecting information about user relationships, the authors gathered also information on tweets and, by performing a semantic analysis, also on the main topics discussed in these tweets. The nal dataset contained 4,262 trending topics and 106 million tweets. From a technical standpoint we want to observe that: (i) The Twitter API allows to access the whole social graph, i.e., the graph representing users and their connections in Twitter, without authentication. Other Social Web platforms and the APIs they oer, however, do not generally allow to access the whole social graph. A meaningful example is given by the Facebook API. (ii) The Twitter API, by default, allows a human user or a software agent to send only 150 requests per hour: this could be an inadmissible limitation because the amount of information generated within Twitter in a relatively small time slot can be very large, and, then, changes in the Twitter network topology could not be properly sensed. To overcome this problem, Twitter oers white lists: users registered to white lists can send up to 20 000 requests per IP per hour. In [Kwak et al. 2010] the authors used a group of 20 computers, each of them belonging to the Twitter white lists, to perform a real-time monitoring of Twitter. Approaches based on the scraping of HTML pages are able to overcome the limitations above even if they are more complicated to design and implement. To the best of our knowledge, one of the rst attempt to crawl large Online Social Networks was performed by Mislove et al. [Mislove et al. 2007]. In that paper, the authors focused on platforms like Orkut, Flickr and LiveJournal. To perform crawling, the approach of [Mislove et al. 2007] suggests to iteratively retrieve the list of friends of a user which have not yet been visited and to add these contact to the list of users to visit. According to the language of the graph theory, this corresponds to perform a Breadth-First-Search (BFS) visit of the Social Network graph. The user account from which the BFS starts is often called seed node; the BFS ends when the whole graph is visited or, alternatively, a stop criterium is met. The BFS is easy to implement and ecient; it produces accurate results if applied on social graphs which can be modeled as unweighted graphs. Due to these reasons, it has been applied in a large number of studies about the topology and structure of Online Social Networks (see, for instance, [Chau et al. 2007; Wilson et al. 2009; Gjoka et al. 2010; Ye et al. 2010; Catanese et al. 2010]). As observed by [Mislove et al. 2007], BFS may incur in heavy limitations. First of all, a crawler can get trapped into a strong connected component of the social graph. In addition, if we would use the BFS sample to estimate some structural properties of the social network graph, some properties could be overestimated while others could be underestimated [Kurant et al. 2010]. To alleviate these problems, several authors suggested more rened sampling techniques. The implementation of these techniques is equivalent to dene new Web Data Extraction procedures. Most of these techniques have been discussed and exploited in the context of Facebook [Gjoka et al. 2010; Catanese et al. 2010; Catanese et al. 2011] but, unfortunately, some of them can not be extended to other platforms.
ACM Computing Surveys, Vol. V, No. N, July 2012.

40

Emilio Ferrara et al.

In particular, Gjoka et al. [Gjoka et al. 2010] considered dierent visiting algorithms, like BFS, Random Walks and Metropolis-Hastings Random Walks. A particular mention goes to a visiting method called rejection sampling. This technique relies on the fact that a truly uniform sample, of Facebook users can be obtained by generating uniformly at random a 32-bit user ID and, subsequently, by polling Facebook about the existence of that ID. The correctness of this procedure derives from the fact that each Facebook user was uniquely identied by a numerical ID ranging from 0 and 232 1. Of course, such a solution works well for Facebook but it could not work for other platforms. In [Catanese et al. 2010; Catanese et al. 2011], the authors designed a Web Data Extraction architecture based on Intelligent Agents, which is reproduced in Figure 10. Such an architecture consists of three main components: (i) a server running the mining agent(s); (ii) a cross-platform Java application, which implements the logic of the agent; (iii) an Apache interface, which manages the information transfer through the Web. The proposed architecture is able to implement several crawling strategies like BFS or rejection sampling. The sampling procedure in [Catanese et al. 2010; Catanese et al. 2011] works as follows: an agent is activated and it queries the Facebook server(s) to obtain the list of Web pages representing the list of friends of a Facebook user. Of course, the Facebook account to visit depends on the basis of the crawling algorithm we want to implement. After parsing this list of pages, it is possible to reconstruct a portion of the Facebook network. Collected data can be converted in XML format in such a way as to they can be exploited by other applications (e.g., network visualization tools).

Fig. 10.

Architecture of the data extraction platform proposed in [Catanese et al. 2011]

Unfortunately, some technical limitations limit the eectiveness of the process described above. In particular, Facebook provides shortened friend-lists not exceeding 400 Facebook members to reduce trac through its network. To avoid these limitations, other techniques like visual data extraction techniques have to be considered. In addition to gathering data about social relationship, we may collect contents generated by users. These contents may vary from resources posted by users (like photos in Flickr or videos in YouTube) to tags applied for labeling resources with the goal of making the retrieval of these resources easier or to increase the visibility of the contents they generated. Extracting contents from a Social Web platform do not pose meaningful technical problems but it should be faster and easier than in other elds: HTML-aware and model-based extraction systems should t very well with the semi-structured
ACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

41

templates used by most common Social Networking services. Sometimes data are distributed under structured formats like RSS/Atom so acquiring this information is easier than with traditional HTML sources. Privacy pitfalls in collecting user related data. The major concern about the extraction of data from Social Web platforms is about privacy. Several researchers, in fact, showed that we can disclose private information about users by leveraging on publicly available information available in a Social Web platform. For instance, [Hecht et al. 2011] provided an approach to nding user location on the basis of user tweets. The authors used basic machine learning techniques which combined user tweets with geotagged articles in Wikipedia. In an analogous fashion, [Kinsella et al. 2011] used geographic coordinates extracted from geotagged Twitter data to model user location. Locations were modeled at dierent levels of granularity ranging from the zip code to the country level. Experimental studies in [Kinsella et al. 2011] show that the proposed model can predict country, state and city with an accuracy comparable with that achieved by industrial tools for geo-localization. Crandall et al. [Crandall et al. 2009] investigated how to organize a large collection of geotagged photos extracted from Flickr. The proposed approach combined content analysis (based on the textual tags typed by users to describe photos), image analysis and structural analysis based on geotagged data. The most relevant result is that it is possible to locate Flickr photos with a high precision by identifying landmarks via visual, temporal and textual features. In [Chaabane et al. 2012] the authors considers a range of user signals expressing user interests (e.g., I like declaration in Facebook) which are often disclosed by users. By means of a semantic-driven inference technique based on an ontologized version of Wikipedia, the authors show how to discover hidden information about users like gender, relationship status and age. From the discussion above it emerges that collecting seemingly innocuous data from a Social Web platform hides high privacy risks. Therefore, collected data should be manipulated in such a way as to reduce these risks. 4.2.2 Extracting data from multiple Online Social Web platforms. Social Web users often create and maintain dierent proles in dierent platforms with dierent goals (e.g., to post and share their photos in Flickr, to share bookmarks on Delicious, to be aware on job proposals on LinkedIn and so on). We rst discuss technical approaches to collecting data from multiple Social Web platforms as well as the opportunities coming from the availability of these data. We nally describe potential privacy risks related to the management of gathered data. Collecting data from multiple Social Web platforms. The main technical challenges encountered by Web Data Extraction techniques to collect data from multiple Social Web platforms consists of linking information referring to the same user or the same object. Early approaches were based on ad-hoc techniques; subsequent approaches featured a higher level of automation and were based on ad-hoc APIs like the Google Social Graph API. To the best of our knowledge, one of the rst approaches facing the problem of correlating multiple user accounts was presented in [Szomszor et al. 2008]. In
ACM Computing Surveys, Vol. V, No. N, July 2012.

42

Emilio Ferrara et al.

that paper, the authors started with a list of 667, 141 user accounts on Delicious such that each account was uniquely associated with a Delicious prole. The same procedure was repeated on a data sample extracted from Flickr. The rst stage in the correlation process consisted in comparing usernames in Delicious and Flickr: if these strings exactly match, the two accounts were considered as referring to the same person. In this way, it was possible to build a candidate list consisting of 232, 391 usernames such that each user name referred to a Flickr and Delicious prole. Of course, such a list must be rened because, for instance, dierent users may choose the same username. Since both in Delicious and Flickr the users had the chance of lling a form by specifying their real names, the authors of [Szomszor et al. 2008] suggested to rene the candidate list by keeping only those user accounts whose real names matched exactly. Such a procedure signicantly lowered the risk of incurring in false associations but at the same time, dramatically reduced the size of the candidate list to only 502 elements. Some tricks aiming at producing a larger and accurate dataset were also proposed in [Szomszor et al. 2008]: in particular, the authors observed that, in real scenarios, if a user creates accounts in several Web sites, she frequently adds a link to her accounts: so by using traditional search engines like Google, we can nd all pages linking to the homepage of a user and, by ltering these hits we can nd the exact URL of the prole of a user in dierent platforms. The process described above can be automatized by exploiting ad hoc tool. Among these tools, the most popular was perhaps the Google Social Graph API, even if such an API is no longer available. Such an API is able to nd connections among persons on the Web. It can be queried through an HTTP request having a URL called node as its parameter. The node species the URL of a Web page of a user u. The Google Social Graph API is able to return two kinds of results: A list of public URLs that are associated with u; for instance, it reveals the URLs of the blog of u and of her Twitter page. A list of publicly declared connections among users. For instance, it returns the list of persons who, in at least one social network, have a link to a page which can be associated with u. The Google Social Graph API has been used in [Abel et al. 2012]. The identication of the connections between persons or Web objects (like photos or videos) is the key step to design advance services often capable of oering a high level of personalization. For instance, in [Shi et al. 2011], the authors suggest to use tags present in different Social Web systems to establish links between items located in each system. In [Abel et al. 2012] the system Mypes is presented. Mypes supports the linkage, aggregation, alignment and semantic enrichment of user proles available in various Social Web systems, such as Flickr, Delicious and Facebook. In the eld of Recommender Systems, the approach of [De Meo et al. 2009] show how to merge ratings provided by users in dierent Social Web platforms with the goal of computing reputation values which are subsequently used to generate recommendations. Some authors have also proposed to combine data coming from dierent platforms but referring to the same object. A nice example has been provided in [Stewart
ACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

43

et al. 2009]; in that paper the authors consider users of blogs concerning music and users of Last.fm, a popular folksonomy whose resources are musical tracks. The ultimate goal of [Stewart et al. 2009] is to enrich each Social Web system by re-using tags already exploited in other environments. This activity has a twofold eect: it rst allows the automatic annotation of resources which were not originally labeled and, then, enriches user proles in such a way that user similarities can be computed in a more precise way. Privacy risks related to the management of user data spread in multiple platforms. The discussion above show that the combination and linkage of data spread in independent Social Web platforms provide clear advantages in term of item description as well as on the level of personalization that a system is able to provide to its subscribers. However, the other side of the coin is given by the privacy problems we may incur when we try to glue together data residing on dierent platforms. Some authors recently introduced the so-called user identication problem, i.e., they studied what information is useful to disclose links between the multiple accounts of a user in independent Social Web platforms. One of the rst approaches to dealing with the user identication problem was proposed in [Vosecky et al. 2009]. In that paper the authors focused on Facebook and StudiVZ21 and investigated which prole attributes can be used to identify users. In [Zafarani and Liu 2009], the authors studied 12 dierent Social Web systems (like Delicious, Flickr and YouTube) with the goal of nding a mapping involving the dierent user accounts. This mapping can be found by applying a traditional search engine. In [Iofciu et al. 2011], the authors suggest to combine prole attributes (like usernames) with an analysis of the user contributed tags to identify users. They suggest various strategies to compare the tag-based proles of two users and some of these strategies were able to achieve an accuracy of almost 80% in user identication. [Perito et al. 2011] explored the possibility of linking users proles only by looking at their usernames. Such an approach is based on the idea that the probability that two usernames are associated with the same person depends on the entropies of the strings representing usernames. [Balduzzi et al. 2010] described a simple but eective attack to discover the multiple digital identities of a user. Such an attack depends on the fact that a user often subscribes to multiple Social Web platforms by means of a single email address. An attacker can query Social Web platforms by issuing a list of e-mail addresses to each Social Web platform; once the proles of a user have been identied in each platform the data contained in these proles can be merged to obtain private personal information. The authors considered a list of about 10.4 millions of e-mail addresses and they were able to automatically identify more than 1.2 millions of user proles registered in dierent platforms like Facebook and XING. The most popular providers acknowledged the privacy problem raised in [Balduzzi et al. 2010] and implemented ad hoc countermeasures. An interesting study is nally presented in [Goga et al. 2012]. The authors
21 http://studivz.net/

ACM Computing Surveys, Vol. V, No. N, July 2012.

44

Emilio Ferrara et al.

showed that correlating seemingly innocuous attributes describing user activities in multiple social networks allow attackers to track user behaviors and infer their features. In particular, the analysis of [Goga et al. 2012] focused on three features of online activity like the geo-location of users posts, the timestamp of posts, and the users writing style (captured by proper language models). 4.3 Opportunities for cross-fertilization

In this section we discuss on the possibility of re-using Web Data Extraction techniques originally developed in a given application domain to another domain. This discussion is instrumental in highlighting techniques which can be applied across dierent application domains and techniques which require some additional information (which can be present in some application domains and missing in other ones). In the most general case, no assumption is made on the structure and content of a collection of Web pages which constitute the input of a Web Data Extraction tool. Each Web page can be regarded as a text document and, in such a conguration, approaches relying on Regular Expressions can be applied independently of the application domain we are considering. As mentioned before, regular expressions can be regarded as a formal language allowing to nd strings or patterns from text according to some matching criteria. Approaches based on regular expressions therefore may be suitable when we do not have, at our disposal, any information about the structure of the Web pages. Of course, these approaches can be extended in such a way as to take some structural elements into account (like HTML tags). The usage of regular expression may be disadvantageous if need to collect large collections of documents and assume that those documents deal with dierent topics. In such a case, we need complex expressions for extracting data from documents and this would require a high level of expertise. The next step consists of assuming that some information on the structure of the document is available. In such a case, wrappers are an eective solution and they are able to work fairly well in dierent application domains. In fact, the hierarchical structure induced by HTML tags associated with a page often provides useful information to the Web Data Extraction task. A powerful solution taking advantage of the structure of HTML pages derives from the the usage of XPath to quickly locate an element (or multiple instances of the same element in the document tree). Approaches based on XPath have been rst exploited in the context of Enterprise applications and, later, they have been successfully re-used in the context of Social Web applications. There are, however, a number of requirements an application domain should satisfy in order to enable the usage of XPath. First of all, the structure of Web pages should be perfectly dened to the wrapper designer (or to the procedure inducing the wrapper); such an assumption, of course, can not be true in some domains because data are regarded as proprietary and technical details about the structure of the Web wrappers in that domain are not available. In addition, we should assume a certain level of structural coherence among all the Web pages belonging to a Web source. Such an assumption is often true both in enterprise domain and in many Social Web systems: in fact, Web pages within an organization (e.g., a rm) or a Social Web platform often derive from the same template and, therefore, a certain form of structural regularity across all the pages
ACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

45

Fig. 11. Facebook Crawling by means of a visual extraction approach. The goal of the wrapper is to extract the names of the friends of a user (highlighted in the gure by means of red frames).

emerge. By contrast, if we plan to manage pages with dierent structures (even if referring to the same domain), we observe that even small changes in the structure of two Web pages may have a devastating impact and thus would require to entirely rewrite the wrapper. A nice example of cross-fertilization is in the context of the crawling of Online Social Networks [Catanese et al. 2011; Catanese et al. 2010]. In that paper the authors implemented a Web wrapper to crawl Facebook largely exploiting techniques and algorithms which were part of the Lixto suite and that were originally designed to work for Business and Competitive Intelligence applications. The proposed wrapper is based on the execution of XPath queries; human experts, in the conguration phase, are allowed to specify what elements have to be selected. The crawler provides two running modes: (i) visual extraction and, (ii) HTTP request-based extraction. In the visual extraction mode, depicted in Figure 11, the crawler embeds a Firefox browser interfaced through XPCOM22 and XULRunner23 . The visual approach requires the rendering of the Web pages which is a timeconsuming activity. Therefore, to extract large amounts of data, [Catanese et al. 2010] suggested to send HTTP requests to fetch Web pages. In some application scenarios, in addition to assuming that the syntactic structure of a Web page is known, it is possible to assume that a rich semantic structure emerges from the Web pages. If such an hypothesis holds true, techniques from Information Extraction (IE) and Natural Language Processing (NLP) can be conveniently used [Winograd 1972; Berger et al. 1996; Manning and Schtze 1999]. The u range of applications beneting from NLP techniques comprises relevant examples both in the enterprise and Social Web scenarios: for instance, relevant applications of NLP/IE techniques are the extraction of facts from speech transcriptions in forums, email messages, newspaper articles, resumes etc.

22 https://developer.mozilla.org/en/xpcom 23 https://developer.mozilla.org/en/XULRunner

ACM Computing Surveys, Vol. V, No. N, July 2012.

46

Emilio Ferrara et al.

5.

CONCLUSIONS AND ONGOING WORK

The World Wide Web contains a large amount of unstructured data. The need for structured information urged researchers to develop and implement various strategies to accomplish the task of automatically extracting data from Web sources. Such a process is known with the name of Web Data Extraction and it has had (and continues to have) a wide range of applications in several elds, ranging from commercial to Social Web applications. The central thread of this survey is to classify existing approaches to Web Data Extraction in terms of the applications for which they have been employed. In the rst part of this paper, we provided a classication of algorithmic techniques exploited to extract data from Web pages. We organized the material by presenting rst basic techniques and, subsequently, the main variants to these techniques. Finally, we focus on how Web Data Extraction systems work in practices. We provide dierent perspectives to classify Web Data Extraction systems (like the ease of use, the possibility of extracting data from the Deep Web and so on). The second part of the survey is about the applications of Web Data Extraction systems to real-world scenarios. We provided a simple classication framework in which existing applications have been grouped into two main classes (Enterprise and Social Web applications). The material has been organized around the application domains of Web Data Extraction systems: we identify, for each class, some subdomains and described how Web Data Extraction techniques work in each domain. This part ends with a discussion about the opportunities of cross-fertilization. 5.1 A glance on the future

In this section we discuss some recent applications of Web Data Extraction techniques, whose importance will certainly increase in the future. A special emphasis is due to the application of Web Data Extraction techniques in the elds of Bioinformatics and Scientic Computing (see Section 5.1), Web Harvesting (see Section 5.1) and, nally, on applications devoted to link data sources on the Web to improve the eectiveness of other applications (see Section 5.1). Bio-informatics and Scientic Computing. A growing eld of application of the Web Data Extraction is bio-informatics: on the World Wide Web it is very common to nd medical sources, in particular regarding bio-chemistry and genetics. BioInformatics is an excellent example of the application of scientic computing refer e.g. to [Descher et al. 2009] for a selected scientic computing project. Plake et al. [Plake et al. 2006] worked on PubMed24 - the biggest repository of medical-scientic works that covers a broad range of topics - extracting information and relationships to create a graph; this structure could be a good starting point to proceed in extracting data about proteins and genes, for example connections and interactions among them: this information can be usually found, not in Web pages, rather they are available in PDF or Postscript format. In the future, Web Data Extraction should be extensively used also to these documents: approaches to solving this problem are going to be developed, inherited, both from Information Extraction and Web Data Extraction systems, because of the semi-structured for24 http://www.pubmed.com/

ACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

47

mat of PostScript-based les. On the other hand, web services play a dominant role in this area as well, and another important challenge is the intelligent and ecient querying of Web services as investigated by the ongoing SeCo project 25 . Web harvesting. One of the most attractive future applications of the Web Data Extraction is Web Harvesting [Weikum 2009]: Gatterbauer [Gatterbauer 2009] denes it as the process of gathering and integrating data from various heterogeneous Web sources. The most important aspect (although partially dierent from specic Web Data Extraction) is that, during the last phase of data transformation, the amount of gathered data is many times greater than the extracted data. The work of ltering and rening information from Web sources ensures that extracted data lie in the domain of interest and are relevant to users: this step is called integration. The Web harvesting remains an open problem with large margin of improvement: because of the billions of Web pages, it is a computational problem, also for restricted domains, to crawl enough sources from the Web to build a solid ontological base. There is also a human engagement problem, correlated to the degree of automation of the process: when and where humans have to interact with the system of Web harvesting? Should be a fully automatic process? What degree of precision can we accept for the harvesting? All these questions are still open for future works. Projects such as the DIADEM 26 at Oxford University tackle the challenge for fully automatic generation of wrappers for restricted domains such as real estate. Linking Data from several Web sources. A growing number of authors suggest to integrate data coming from as many sources as possible to obtain a detailed description of an object. Such a description would be hard to obtain if we would focus only on a single system/service. The material discussed in this section partially overlaps with some ideas/techniques presented in Section 4.2.2. However, in Section 4.2.2 we focused on Social Web platforms and showed that linking information stored in the proles of a user spread across multiple platforms leads to a better identication of user needs and, ultimately, raises the level of personalization a system can oer to her. In this section, by contrast, we focus on systems having two main features. (i) they often publicly expose their data on the Web and (ii) these systems are lacking of a social connotation, i.e., they do not target at building a community of interacting members. A rst research work showing the benets of linking data provided by independent Web systems is provided in [Szomszor et al. 2007]. In that paper, the authors combined information from the Internet Movie Database www.imdb.com) and Netix (www.netflix.com). The IMDB is an online database containing extensive information on movies, actors and television shows. IMDB users are allowed to add tags to describe the main features of a movie (e.g., the most important scenes, location, genres and so on). Netix is an US-based company oering an online DVD rental service. Netix users are allowed to rate a movie by providing a score. In [Szomszor et al. 2007], data from Netix and IMDB were imported in a relational DBMS; movie titles in IMDB were correlated with movie titles in Netix by ap25 http://www.search-computing.it/ 26 http://web.comlab.ox.ac.uk/projects/DIADEM/

ACM Computing Surveys, Vol. V, No. N, July 2012.

48

Emilio Ferrara et al.

plying string matching techniques. In this way, for each movie, the Netix ratings and the IMDB description were available. The authors studied three recommendation strategies based on ratings alone, on tags alone and, nally, a combination of ratings and tags. Experimental trials provided evidence that the combination of data located in dierent systems was able to improve the level of accuracy in the provided recommendations. Spiegel et al. presented an analogous study in [Spiegel et al. 2009]; in that paper the authors combine user ratings (coming from MovieLens database) with movie information provided by IMDB. The linkage of datasets coming from independent Web platforms fuels novel scientic applications. One of these application is given from cross-domain recommender systems, i.e., on recommender systems running on multiple domains. For example, one can use user ratings about movies to predict user preferences in the music domain. In a given domain (e.g., if our main aim is to recommend movies), information about user preferences may be insucient and, in many case, most of this information is missing for a non-negligible part of user population. By contrast, a relevant amount of information describing user preferences could be available (at a limited cost) in other domains. We could transfer this information from a domain to another one with the goal of dealing with data sparsity and improving recommendation accuracy. Various strategies have been proposed to perform such a transfer of information and some of these strategies take advantages from models and algorithms developed in the eld of transfer of learning [Pan and Yang 2010]. For instance, some approaches use co-clustering [Li et al. 2009a], other clustering techniques in conjunction with probabilistic models [Li et al. 2009b] and, nally, other approaches project the user and item space in the various domains in a shared latent space by means of regularization techniques [Zhang et al. 2010].
REFERENCES Abel, F., Herder, E., Houben, G.-J., Henze, N., and Krause, D. 2012. Cross-system User Modeling and Personalization on the Social Web. User Modeling and User-Adapted Interaction (UMUAI), To Appear. Amalfitano, D., Fasolino, A. R., and Tramontana, P. 2008. Reverse engineering nite state machines from rich internet applications. In WCRE 08: Proc. of the 2008 15th Working Conference on Reverse Engineering. IEEE Computer Society, Washington, DC, USA, 6973. Backstrom, L., Boldi, P., Rosa, M., Ugander, J., and Vigna, S. 2011. Four degrees of separation. Arxiv preprint arXiv:1111.4570 . Balduzzi, M., Platzer, C., Holz, T., Kirda, E., Balzarotti, D., and Kruegel, C. 2010. Abusing social networks for automated user proling. In Proc. of the International Symposium on Recent Advances in Intrusion Detection (RAID 2010). Lecture Notes in Computer Science. Springer, Ottawa, Ontario, Canada, 422441. Baumgartner, R., Campi, A., Gottlob, G., and Herzog, M. 2010. Web data extraction for service creation. Search Computing: Challenges and Directions. Baumgartner, R., Ceresna, M., and Ledermuller, G. 2005. Deepweb navigation in web data extraction. In CIMCA 05: Proc. of the International Conference on Computational Intelligence for Modelling, Control and Automation and International Conference on Intelligent Agents, Web Technologies and Internet Commerce Vol-2 (CIMCA-IAWTIC06). IEEE Computer Society, Washington, DC, USA, 698703. Baumgartner, R., Flesca, S., and Gottlob, G. 2001a. The elog web extraction language. In LPAR 01: Proc. of the Articial Intelligence on Logic for Programming. Springer-Verlag, London, UK, 548560.
ACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

49

Baumgartner, R., Flesca, S., and Gottlob, G. 2001b. Visual web information extraction with lixto. In VLDB 01: Proc. of the 27th International Conference on Very Large Data Bases. Morgan Kaufmann Publishers Inc., San Francisco, CA, USA, 119128. Baumgartner, R., Frolich, O., Gottlob, G., Harz, P., Herzog, M., Lehmann, P., and Wien, T. 2005. Web data extraction for business intelligence: the lixto approach. In Proc. of BTW 2005. 4865. Baumgartner, R., Froschl, K., Hronsky, M., Pottler, M., and Walchhofer, N. 2010. Semantic online tourism market monitoring. Proc. of the 17th ENTER eTourism International Conference. Baumgartner, R., Gatterbauer, W., and Gottlob, G. 2009. Web data extraction system. Encyclopedia of Database Systems, 34653471. Baumgartner, R., Gottlob, G., and Herzog, M. 2009. Scalable web data extraction for online market intelligence. Proc. of the International Conference on Very Large Databases 2, 2, 1512 1523. Berger, A. L., Pietra, V. J. D., and Pietra, S. A. D. 1996. A maximum entropy approach to natural language processing. Comput. Linguist. 22, 1, 3971. Berthold, M. and Hand, D. J. 1999. Intelligent Data Analysis: An Introduction. SpringerVerlag New York, Inc., Secaucus, NJ, USA. Bettencourt, L., Cintron-Arias, A., Kaiser, D., and Castillo-Chavez, C. 2006. The power of a good idea: Quantitative modeling of the spread of ideas from epidemiological models. Physica A: Statistical Mechanics and its Applications 364, 513536. Catanese, S., De Meo, P., Ferrara, E., and Fiumara, G. 2010. Analyzing the facebook friendship graph. In Proc. of the 1st International Workshop on Mining the Future Internet. 1419. Catanese, S., De Meo, P., Ferrara, E., Fiumara, G., and Provetti, A. 2011. Crawling facebook for social network analysis purposes. In Proc. of the International Conference on Web Intelligence, Mining and Semantics (WIMS 2011). ACM, Sogndal, Norway, 52. Chaabane, A., cs, G., and Kaafar, M. 2012. You Are What You Like! Information Leakage Through Users Interests. In Proc. of the Annual Network and Distributed System Security Symposium (NDSS 2012). Chang, C.-H., Kayed, M., Girgis, M. R., and Shaalan, K. F. 2006. A survey of web information extraction systems. IEEE Trans. on Knowl. and Data Eng. 18, 10, 14111428. Chau, D., Pandit, S., Wang, S., and Faloutsos, C. 2007. Parallel crawling for online social networks. In Proceedings of the 16th International Conference on the World Wide Web. 1283 1284. Chen, F., Doan, A., Yang, J., and Ramakrishnan, R. 2008. Ecient information extraction over evolving text data. In Proc. of the IEEE 24th International Conference on Data Engineering (ICDE 2008). IEEE, Cancun, Mexico, 943952. Chen, H., Chau, M., and Zeng, D. 2002. Ci spider: a tool for competitive intelligence on the web. Decis. Support Syst. 34, 1, 117. Chen, W. 2001. New algorithm for ordered tree-to-tree correction problem. Journal of Algorithms 40, 2, 135158. Collins, M. 1996. A new statistical parser based on bigram lexical dependencies. In Proceedings of the 34th annual meeting on Association for Computational Linguistics. Association for Computational Linguistics, 184191. Crandall, D., Backstrom, L., Huttenlocher, D., and Kleinberg, J. 2009. Mapping the worlds photos. In Proc. of the International Conference on World Wide Web (WWW 2009). ACM, Madrid, Spain, 761770. Crescenzi, V. and Mecca, G. 2004. Automatic information extraction from large websites. J. ACM 51, 5, 731779. Crescenzi, V., Mecca, G., and Merialdo, P. 2001. Roadrunner: Towards automatic data extraction from large web sites. In VLDB 01: Proc. of the 27th International Conference on Very Large Data Bases. Morgan Kaufmann Publishers Inc., San Francisco, CA, USA, 109118.
ACM Computing Surveys, Vol. V, No. N, July 2012.

50

Emilio Ferrara et al.

Crescenzi, V., Mecca, G., and Merialdo, P. 2004. Improving the expressiveness of roadrunner. In SEBD. 6269. Dave, K., Lawrence, S., and Pennock, D. M. 2003. Mining the peanut gallery: opinion extraction and semantic classication of product reviews. In WWW 03: Proc. of the 12th international conference on World Wide Web. ACM, New York, NY, USA, 519528. De Meo, P., Nocera, A., Quattrone, G., Rosaci, D., and Ursino, D. 2009. Finding reliable users and social networks in a social internetworking system. In Proc. of the International Database Engineering and Applications Symposium (IDEAS 2009). ACM Press, Cetraro, Cosenza, Italy, 173181. Descher, M., Feilhauer, T., Ludescher, T., Masser, P., Wenzel, B., Brezany, P., Elsayed, I., Wohrer, A., Tjoa, A. M., and Huemer, D. 2009. Position paper: Secure infrastructure for scientic data life cycle management. In ARES. 606611. Fayzrakhmanov, R., Goebel, M., Holzinger, W., Kruepl, B., Mager, A., and Baumgartner, R. 2010. Modelling web navigation with the user in mind. In Proc. of the 7th International Cross-Disciplinary Conference on Web Accessibility. Ferrara, E. and Baumgartner, R. 2011a. Automatic wrapper adaptation by tree edit distance matching. Combinations of Intelligent Methods and Applications, 4154. Ferrara, E. and Baumgartner, R. 2011b. Design of automatically adaptable web wrappers. In Proceedings of the 3rd International Conference on Agents and Articial Intelligence. 211217. Ferrara, E. and Baumgartner, R. 2011c. Intelligent self-repairable web wrappers. Lecture Notes in Computer Science 6934, 274285. Fiumara, G. 2007. Automated information extraction from web sources: a survey. Proc. of Between Ontologies and Folksonomies Workshop in 3rd International Conference on Communities and Technology, 19. Flesca, S., Manco, G., Masciari, E., Rende, E., and Tagarelli, A. 2004. Web wrapper induction: a brief survey. AI Commun. 17, 2, 5761. Furche, T., Gottlob, G., Grasso, G., Gunes, O., Guo, X., Kravchenko, A., Orsi, G., Schallhart, C., Sellers, A. J., and Wang, C. 2012. Diadem: domain-centric, intelligent, automated data extraction methodology. In WWW (Companion Volume). 267270. Furche, T., Gottlob, G., Grasso, G., Schallhart, C., and Sellers, A. J. 2011. Oxpath: A language for scalable, memory-ecient data extraction from web applications. PVLDB 4, 11, 10161027. Garrett, J. J. 2005. Ajax: A new approach to web applications. Tech. rep., Adaptive Path. Gatterbauer, W. 2009. Web harvesting. Encyclopedia of Database Systems, 34723473. Gatterbauer, W. and Bohunsky, P. 2006. Table extraction using spatial reasoning on the css2 visual box model. In AAAI 06 Proc. of the 21st national conference on Articial intelligence. AAAI Press, 13131318. Gjoka, M., Kurant, M., Butts, C., and Markopoulou, A. 2010. Walking in Facebook: a case study of unbiased sampling of OSNs. In Proc. of the 29th Conference on Information Communications. IEEE, 24982506. Goga, O., Lei, H., Parthasarathi, S., Friedland, G., Sommer, R., and Teixeira, R. 2012. On exploiting innocuous user activity for correlating accounts across social network sites. Tech. rep., ICSI Technical Reports - University of Berkeley. Gottlob, G. and Koch, C. 2004a. Logic-based web information extraction. SIGMOD Rec. 33, 2, 8794. Gottlob, G. and Koch, C. 2004b. Monadic datalog and the expressive power of languages for web information extraction. J. ACM 51, 1, 74113. Hammersley, B. 2005. Developing feeds with rss and atom. OReilly. Han, J. and Kamber, M. 2000. Data mining: concepts and techniques. Morgan Kaufmann Publishers Inc., San Francisco, CA, USA. Hecht, B., Hong, L., Suh, B., and Chi, E. 2011. Tweets from Justin Biebers heart: the dynamics of the location eld in user proles. In Proc. of the International Conference on Human Factors in Computing Systems (CHI 2011). ACM, Vancouver, British Columbia, Canada, 237246.
ACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

51

Hsu, C.-N. and Dung, M.-T. 1998. Generating nite-state transducers for semi-structured data extraction from the web. Inf. Syst. 23, 9, 521538. Huck, G., Fankhauser, P., Aberer, K., and Neuhold, E. 1998. JEDI: Extracting and synthesizing information from the web. In Proc. of COOPIS. Iofciu, T., Fankhauser, P., Abel, F., and Bischoff, K. 2011. Identifying Users Across Social Tagging Systems. In Proc. of the International Conference on Weblogs and Social Media (ICWSM 11). Barcelona, Spain. Irmak, U. and Suel, T. 2006. Interactive wrapper generation with minimal user eort. In Proc. of the International Conference on World Wide Web (WWW 2006)). ACM, Edinburgh, Scotland, 553563. Kaiser, K. and Miksch, S. 2005. Information extraction. a survey. Tech. rep., E188 - Institut f r Softwaretechnik und Interaktive Systeme; Technische Universitt Wien. u a Khare, R. and Celik, T. 2006. Microformats: a pragmatic path to the semantic web. In WWW 06: Proc. of the 15th international conference on World Wide Web. ACM, New York, NY, USA, 865866. Kim, Y., Park, J., Kim, T., and Choi, J. 2007. Web information extraction by html tree edit distance matching. In International Conference on Convergence Information Technology. IEEE, 24552460. Kinsella, S., Murdock, V., and OHare, N. 2011. Im eating a sandwich in Glasgow: modeling locations with tweets. In Proc. of the International Workshop on Search and Mining usergenerated contents. ACM, Glasgow, UK, 6168. Kleinberg, J. 2000. The small-world phenomenon: an algorithm perspective. In Proc. of the ACM symposium on Theory of Computing (STOC 2000). ACM, Portland, Oregon, USA, 163 170. Kleinberg, J. 2008. The convergence of social and technological networks. Communications of the ACM 51, 11, 6672. Kohlschutter, C., Fankhauser, P., and Nejdl, W. 2010. Boilerplate detection using shallow text features. In Proceedings of the third ACM international conference on Web search and data mining. ACM, 441450. Krupl, B., Herzog, M., and Gatterbauer, W. 2005. Using visual cues for extraction of tabular data from arbitrary html documents. In WWW 05: Special interest tracks and posters of the 14th international conference on World Wide Web. ACM, New York, NY, USA, 10001001. Krupl-Sypien, B., Fayzrakhmanov, R. R., Holzinger, W., Panzenbock, M., and Baumgartner, R. 2011. A versatile model for web page representation, information extraction and content re-packaging. In ACM Symposium on Document Engineering. 129138. Kurant, M., Markopoulou, A., and Thiran, P. 2010. On the bias of breadth rst search (bfs) and of other graph sampling techniques. In Proceedings of the 22nd International Teletrac Congress. 18. Kushmerick, N. 1997. Wrapper induction for information extraction. Ph.D. thesis, University of Washington. Chairperson-Weld, Daniel S. Kushmerick, N. 2000. Wrapper induction: eciency and expressiveness. Artif. Intell. 118, 1-2, 1568. Kushmerick, N. 2002. Finite-state approaches to web information extraction. Proc. of 3rd Summer Convention on Information Extraction, 7791. Kwak, H., Lee, C., Park, H., and Moon, S. 2010. What is Twitter, a social network or a news media? In Proc. of the International Conference on World Wide Web (WWW 2010). ACM, Raleigh, North Carolina, USA, 591600. Laender, A. H. F., Ribeiro-Neto, B. A., da Silva, A. S., and Teixeira, J. S. 2002. A brief survey of web data extraction tools. SIGMOD Rec. 31, 2, 8493. Li, B., Yang, Q., and Xue, X. 2009a. Can movies and books collaborate? cross-domain collaborative ltering for sparsity reduction. In Proc. of the International Joint Conference on Articial Intelligence (IJCAI 2009). Morgan Kaufmann Publishers Inc., Innsbruck, Austria, 20522057.
ACM Computing Surveys, Vol. V, No. N, July 2012.

52

Emilio Ferrara et al.

Li, B., Yang, Q., and Xue, X. 2009b. Transfer learning for collaborative ltering via a ratingmatrix generative model. In Proc. of the International Conference on Machine Learning (ICML 2009). Montreal, Canada, 617624. Liu, B. 2011. Structured data extraction: Wrapper generation. Web Data Mining, 363423. Liu, L., Pu, C., and Han, W. 2000. XWrap: An extensible wrapper construction system for internet information. In Proc. of ICDE. Manning, C. D. and Schutze, H. 1999. Foundations of statistical natural language processing. MIT Press, Cambridge, MA, USA. ` Masanes, J. 2006. Web archiving: issues and methods. Web Archiving, 153. Mathes, A. 2004. Folksonomies-cooperative classication and communication through shared metadata. Computer Mediated Communication 47, 10. May, W., Himmeroder, R., Lausen, G., and Ludascher, B. 1999. A unied framework for wrapping, mediating and restructuring information from the web. In WWWCM. Sprg. LNCS 1727. Melomed, E., Gorbach, I., Berger, A., and Bateman, P. 2006. Microsoft SQL Server 2005 Analysis Services (SQL Server Series). Sams, Indianapolis, IN, USA. Meng, X., Hu, D., and Li, C. 2003. Schema-guided wrapper maintenance for web-data extraction. In WIDM 03: Proc. of the 5th ACM international workshop on Web information and data management. ACM, New York, NY, USA, 18. Mislove, A., Marcon, M., Gummadi, K., Druschel, P., and Bhattacharjee, B. 2007. Measurement and analysis of online social networks. In Proceedings of the 7th SIGCOMM. ACM, 2942. Monge, A. E. 2000. Matching algorithm within a duplicate detection system. IEEE Techn. Bulletin Data Engineering 23, 4. Muslea, I., Minton, S., and Knoblock, C. 1999. A hierarchical approach to wrapper induction. In AGENTS 99: Proc. of the third annual conference on Autonomous Agents. ACM, New York, NY, USA, 190197. Newman, M. 2003. The structure and function of complex networks. SIAM review , 167256. Oro, E., Ruffolo, M., and Staab, S. 2010. Sxpath - extending xpath towards spatial querying on web documents. PVLDB 4, 2, 129140. Pan, S. and Yang, Q. 2010. A survey on transfer learning. Knowledge and Data Engineering, IEEE Transactions on 22, 10, 13451359. Perito, D., Castelluccia, C., Kaafar, M., and Manils, P. 2011. How unique and traceable are usernames? In Privacy Enhancing Technologies. Springer, 117. Phan, X., Horiguchi, S., and Ho, T. 2005. Automated data extraction from the web with conditional models. Int. J. Bus. Intell. Data Min. 1, 2, 194209. Plake, C., Schiemann, T., Pankalla, M., Hakenberg, J., and Leser, U. 2006. Alibaba: Pubmed as a graph. Bioinformatics 22, 19, 24442445. Rahm, E. and Bernstein, P. A. 2001. A survey of approaches to automatic schema matching. The VLDB Journal 10, 4, 334350. Rahm, E. and Do, H. H. 2000. Data cleaning: Problems and current approaches. IEEE Bulletin on Data Engineering 23, 4. Romero, D. and Kleinberg, J. 2010. The Directed Closure Process in Hybrid Social-Information Networks, with an Analysis of Link Formation on Twitter. In Proceedings of the 4th International Conference on Weblogs and Social Media. Sahuguet, A. and Azavant, F. 1999. Building light-weight wrappers for legacy web data-sources using w4f. In VLDB 99: Proc. of the 25th International Conference on Very Large Data Bases. Morgan Kaufmann Publishers Inc., San Francisco, CA, USA, 738741. Sarawagi, S. 2008. Information extraction. Found. Trends databases 1, 3, 261377. Schifanella, R., Barrat, A., Cattuto, C., Markines, B., and Menczer, F. 2010. Folks in folksonomies: social link prediction from shared metadata. In Proc. of the third ACM international conference on Web Search and Data Mining (WSDM 2010). ACM Press, New York, USA, 271280.
ACM Computing Surveys, Vol. V, No. N, July 2012.

Web Data Extraction, Applications and Techniques: A Survey

53

Selkow, S. 1977. The tree-to-tree editing problem. Information processing letters 6, 6, 184186. Shi, Y., Larson, M., and Hanjalic, A. 2011. Tags as bridges between domains: Improving recommendation with tag-induced cross-domain collaborative ltering. In Proc. of the International Conference on User Modeling, Adaption and Personalization (UMAP 2011). Lecture Notes in Computer Science. Springer, Girona, Spain, 305316. Spiegel, S., Kunegis, J., and F.Li. 2009. Hydra: a hybrid recommender system [cross-linked rating and content information]. In Proc. of the ACM International Workshop on Complex Networks Meet Information & Knowledge Management (CNIKM 2009). Hong Kong, China, 7580. Stewart, A., Diaz-Aviles, E., Nejdl, W., Marinho, L. B., Nanopoulos, A., and SchmidtThieme, L. 2009. Cross-tagging for personalized open social networking. In Proc. of the ACM International Conference on Hypertext and Hypermedia (HT 2009). Torino, Italy, 271278. Szomszor, M., Cantador, I., and Alani, H. 2008. Correlating user proles from multiple folksonomies. In Proc. of the ACM International Conference on Hypertext and Hypermedia (HT08). ACM, Pittsburgh, PA, USA, 3342. Szomszor, M., Cattuto, C., Alani, H., OHara, K., Baldassarri, A., Loreto, V., and Servedio, V. 2007. Folksonomies, the semantic web, and movie recommendation. In Proc. of the European Semantic Web Conference (ESWC 2007). Innsbruck, Austria. ` Turmo, J., Ageno, A., and Catala, N. 2006. Adaptive information extraction. ACM Comput. Surv. 38, 2, 4. Vosecky, J., Hong, D., and Shen, V. 2009. User identication across multiple social networks. In Proc. of the International Conference on Networked Digital Technologies (NDT 09). IEEE, Ostrava, Czech Republic, 360365. Wang, J., Shapiro, B., Shasha, D., Zhang, K., and Currey, K. 1998. An algorithm for nding the largest approximately common substructures of two trees. Pattern Analysis and Machine Intelligence, IEEE Transactions on 20, 8, 889895. Wang, P., Hawk, W. B., and Tenopir, C. 2000. Users interaction with world wide web resources: an exploratory study using a holistic approach. Inf. Process. Manage. 36, 229251. Weikum, G. 2009. Harvesting, searching, and ranking knowledge on the web: invited talk. In WSDM 09: Proc. of the Second ACM International Conference on Web Search and Data Mining. ACM, New York, NY, USA, 34. Wilson, C., Boe, B., Sala, A., Puttaswamy, K., and Zhao, B. 2009. User interactions in social networks and their implications. In Proceedings of the 4th European Conference on Computer Systems. ACM, 205218. Winkler, W. 1999. The state of record linkage and current research problems. In Statistical Research Division, US Census Bureau. Winograd, T. 1972. Understanding Natural Language. Academic Press, Inc., Orlando, FL, USA. Yang, W. 1991. Identifying syntactic dierences between two programs. Software - Practice and Experience 21, 7, 739755. Ye, S., Lang, J., and Wu, F. 2010. Crawling Online Social Graphs. In Proceedings of the 12th International Asia-Pacic Web Conference. IEEE, 236242. Zafarani, R. and Liu, H. 2009. Connecting corresponding identities across communities. In Proc. of the International Conference on Weblogs and Social Media (ICWSM 09). The AAAI Press, San Jose, California, USA, 354357. Zanasi, A. 1998. Competitive intelligence through data mining public sources. In Competitive intelligence review. Vol. 9. Wiley, New York, NY, ETATS-UNIS (1990-2001) (Revue), 4454. Zhai, Y. and Liu, B. 2005. Web data extraction based on partial tree alignment. In WWW 05: Proc. of the 14th international conference on World Wide Web. ACM, NY, 7685. Zhai, Y. and Liu, B. 2006. Structured data extraction from the web based on partial tree alignment. IEEE Trans. on Knowl. and Data Eng. 18, 12, 16141628. Zhang, K., Statman, R., and Shasha, D. 1992. On the editing distance between unordered labeled trees. Information processing letters 42, 3, 133139.
ACM Computing Surveys, Vol. V, No. N, July 2012.

54

Emilio Ferrara et al.

Zhang, Y., Cao, B., and Yeung, D. 2010. Multi-domain collaborative ltering. In Proc. of the 26th Conference on Uncertainty in Articial Intelligence (UAI). AUAI Press, Catalina Island, California, USA, 725732. Zhang, Z., Zhang, C., Lin, Z., and Xiao, B. 2010. Blog extraction with template-independent wrapper. In Network Infrastructure and Digital Content, 2010 2nd IEEE International Conference on. IEEE, 313317. Zhao, H. 2007. Automatic wrapper generation for the extraction of search result records from search engines. Ph.D. thesis, State University of New York at Binghamton, Binghamton, NY, USA. Adviser-Meng, Weiyi.

Received: June 2010; revised: July 2012;

ACM Computing Surveys, Vol. V, No. N, July 2012.