© 2015 IJSRSET | Volume 1 | Issue 2 | Print ISSN : 2395-1990 | Online ISSN : 2394-4099

Themed Section: Engineering and Technology

Fast Searching of Nearest Neighbor using Key Values in Data Mining

Sri Vidhya. A, Prof. Ashwin. M
Dept of CSE, Adhiyamaan College of engineering, Hosur, Tamilnadu, India

ABSTRACT

Spatial query which focus only on the geometrics properties of an object like points, rectangle etc. Now a day‟s
many new applications which involve the queries that completely aim to return an object which satisfy equally on
spatial predicate and their associated text. Spatial query takes the given location and a keyword as the input and
finds the object that matches the both spatial predicate and the text related to the given query. Some of the spatial
queries are range search and nearest neighbor retrieval which includes only geometric properties of an object. For
example, In case of considering all the hotels, a nearest neighbor query would find for the hotel which is near, along
with menu that user required to have in hotel among all the hotels in particular location simultaneously. At present
the better solution is based on IR2-Tree which as few drawbacks that affect the efficiency in query retrieval. So we
develop a new method Spatial inverted index that cope with 3D data to answer the nearest neighbor query using
keyword along with key values in real time. Searching nearest neighbor query using key values will result in quick
response of query when compared to keyword in real time.
Keywords: Spatial Query, Nearest Neighbour Search, IR2-Tree, Key Value and Spatial inverted index

I. INTRODUCTION nearest neighbour retrieval will discover the restaurant

Knowledge and Data Engineering focus mainly on nearest to a given address.
extraction of data from database. i.e. Data mining .Data
mining helps in extraction of most important data from
the huge collection of data that is stored i.e. data
warehouse. Data warehouse is a subject-oriented,
integrated, time-variant and non-volatile collection of
data in support of management's decision making
process. A spatial database is used to store huge amount
of space related data such as maps, medical imaging data
etc. and manages multidimensional objects (such as
points, rectangles, etc.), and provides quick access to
those objects based on different choice criteria. The
importance of spatial databases [1] is it provides a
Figure 1: (a) shows the location of hotels and (b) given their associated text
convenient path to model the entities of reality in a i.e menu‟s.
geometric manner. As an example, locations of
restaurants, hotels, hospitals and then on are typically Let H be a set of multidimensional points in Figure 1.(a)
shown as points in an exceedingly map, whereas larger which indicates the set of hotels around the query point
extents like parks, lakes, and landscapes typically as a „q‟ i.e. static location. As our aim is to combine keyword
mixture of rectangles. Several functionalities of a spatial search with the existing location-searching services on
information are helpful in numerous ways that in facilities such as hospitals, restaurants, hotels, etc., we
specific contexts. For example, in an exceedingly will focus on 3D data‟s, but our technique can be
geographic data system, vary search may be deployed to extended to arbitrary dimensionalities with no technical
search out all restaurants in a particular space, whereas obstacle. We will assume that the points in H have

IJSRSET151180 | Received: 24 March 2015 | Accepted: 11 March 2015 | March-April 2015 [(1)2: 81-85]
81
integer coordinates, such that each coordinate ranges in facilities, or if H is a hospital, Wh can be the list of its
[0, t], where t is a large integer. This is not as restrictive out-patient specialties. It is clear that Wh may potentially
as it may seem, because even if one would like to insist contain numerous words. Traditional nearest neighbor
on real valued coordinates, the set of different search [3] returns the data point closest to a query point.
coordinates represent able under a space limit is still Following [12], we extend the problem to include
finite and enumerable; therefore, we could as well predicates on objects‟ texts. Formally, in our context, a
convert everything to integer with proper scaling. nearest neighbor (NN) query specifies a point q and a set
Wq of keywords (we refer to Wq as the document of the
Each point in H in Figure 1. (b) is associated with a set query). It returns the point in Hq that is the nearest to q,
of words (h1,h2,h3,h4,h5,h6,h7,h8), and termed the Where Hq is defined as
document of H which is denoted as Wh. For example, if
H stands for a restaurant, Wh can be its menu, or if H is a Hq = {h є H | Wq ⫃ Wh} (1)
hotel, Wh can be the description of its services and

II. LITERATURE REVIEW

TABLE 1: Content of Previous Work in Table

AUTHOR PAPER TITLE CONCEPT APPROACHES

G. Cong, C.S. Jensen, and Efficient Retrieval of the Top- Keyword Computes the relevance
D. Wu[7] k based nearest between the object p and
Most Relevant Spatial Web neighbor query q.It returns Partial
Objects Query satisfaction.
Y.-Y. Chen, T. Suel, and Efficient Query Processing Combines Geographic web search
A. Markowetz[8] in Geographic Web Search keyword returns based on ranking
R. Hariharan, B. Hore, C. Engines.[8] search and
Li, and S. Mehrotra, [13] Processing Spatial- range queries
Keyword (SK) Queries in
Geographic Information
Retrieval (GIR)
Systems[13]
I.D. Felipe, V. Hristidis, Keyword Search on Spatial Nearest IR2-Tree preserves
and N. Rishe,[9] Databases[9] neighbor object‟s spatial proximity,
search using which solve the spatial
keywords by query efficiently.
IR2-Tree Here there is no partial
satisfaction in returning
the output.
D. Zhang, Y.M. Chee, A. Keyword Search in Spatial Search for m- Collaborative in nature,
Mondal, A.K.H. Tung, Databases: Towards closest resulting m points should
and M. Kitsuregawa,[14] Searching by Document keyword cover the query keyword
together.
X. Cao, G. Cong, C.S. Collective Spatial Collective Similar idea as the above
Jensen, and B.C. Ooi,[5] Keyword Querying Spatial but it mainly focuses at
keyword optimizing different
querying objective function.
X. Cao, G. Cong, and Retrieving Top-k Prestige- Prestige based Evaluates the similarity of
C.S. Jensen, [4] Based Relevant Spatial Web spatial an object p to a query q by
Objects keyword considering the objects p
neighbours.

International Journal of Scientific Research in Science, Engineering and Technology (ijsrset.com)

82
 III. METHODS AND MATERIAL equals 01101, while that of f(q, s) equals 01111.Given a
question keyword w, SC performs the
A. Ir2-Tree membership ,take a glance at in W by checking whether
Information Retrieval R-tree which is the state of or not or not all the 1‟s of h(w) appear at constant
answering the NN Queries.IR2-tree is the combination positions at intervals the signature of W.
of the R-tree[2] with signature files. First we will see
about Signature file. Signature file refers to a hash- TABLE 2:- Example of bit string computation with l =5 and
based framework, whose instantiation in [9] is called as m = 2.
superimposed coding (SC), which is shown to be more
effective than other instantiation [5]. It is charted to
perform membership tests that is to determine whether a
query word q exists in a set of Words W[11].If it returns
“no”, then q is surely not in W. If SC returns “yes”, then
q is in W, to avoid a false hit W is scanned as a whole if
it returns “yes”. The IR2-tree is an R-tree where each
leaf or non-leaf entry E is augmented with a signature If not, it‟s secured that w cannot belong to W. Otherwise;
which summarizes the union of the texts in the sub tree. the take a glance at can‟t be resolved
victimization alone the signature, and a scan of W
On traditional R-trees, the best-first algorithm [12] is a
follows. [10]A false hit happens if the scan reveals that
well-known solution to NN search. It is now directly to W really doesn‟t contain w. As associate example,
adapt it to IR2-trees. assume that we tend to would like to envision whether or
B. Working of Signature file not or not word c may be a member of
set mistreatment the set‟s signature 01101. Since the
fourth bit of h(r) = zero11 is one but that of 01101 is 0,
SC in real time reports “no”. As another example,
take under consideration the membership take a
glance at of r in whose signature is 01111. This time, SC
returns “yes” as a results of 01111 has 1‟s within
the least the bits where h(r) is on the brink of 1; as a
result, a full scan of the set is required to verify that this
will be a false hit.

C. Inverted Indexes

The inverted index system is also a central part of a

typical software system categorization formula. A goal
of a research engine performance is to optimize the
speed of the query: Understand the
documents where word happens. Once associate index is
developed, that provisions lists of words per document;
it‟s next inverted to develop associate inverted index
[10]. Querying the index would wish serial iteration
through each document and to each word to verify the
same document. The time, memory and method property
Figure 2: Working of SC to execute such a question don‟t appear to be forever in
theory realistic. Instead of listing the words per
For example, Consider the Table.2 which describes the article inside the index, the inverted index system is
bit string computation with l=5 and m=2.In that the bit developed that lists the documents per word. The
string h(p) of p, the third and fifth (counting from left) inverted index created, the question can presently be
determined by jumping to the word id inside the inverted
bits are set to 1. As mentioned earlier, the bit signature
index. Table.3. describes the inverted index of above
of a set W of words simply ORs the bit strings of all the example i.e for the associated text (p,q,r,s,t).
members of W. For instance, the signature of a set f(p,q)

International Journal of Scientific Research in Science, Engineering and Technology (ijsrset.com)

83
 Table.3. Example of an inverted index Table 4:.Example for data set in location Koyambedu.

D. Spatial inverted index (SI-index).

Spatial inverted index [6] will merge multiple list by

their Id‟s instead we will additionally we provide the R-
trees [2] to browse the purposes of all relevant hotel lists
that near to the static point i.e. particular location. This A. Key Value Implementation
access methodology will incorporates the purpose of
coordinates into a traditional inverted index. In our application, we deal with 10 locations which is
specified above. Admin has the full rights to edit, update
the data set in database. While adding the location to the
I. RESULTS AND DISCUSSION database, a key value is generated for each location in
sequence manner. For example, Table.5. Describes the
The Modules are implemented by using HTML, key value for each location. This key value act as the
JAVA, JAVA Script, JSP, and Database. In the primary key in the database for a particular location.
implementation part, a web application is created
Table.5.Examples for key value assigned to locations.
with the page where the user can post their query
and retrieve their response which satisfies the user
requirements. For example, it would be fairly useful
if a search engine can find the nearest hotel that
offers user menu all at same time but this cannot be
done in case of global. This can achieved locally
within a particular zone with N location by making
one location as static location among the N location.

We create an application with different web pages like

Admin login, Admin data entry, User Registration, User
Login, User Keyword Search etc. This application takes
particular ten location of Chennai City like Koyambedu,
T-Nagar, Avadi, Adyar, Guindy, Numgambakkam,
Ambattur, Poonamalee, Anna Nagar, and Ashok Nagar.

Data Set of Each location are collected and stored B. Building R-Tree
in database which includes the Hotel And its details R-Tree is a real tree which is the graphical
like name, address ,Specialize , phone number and representation of the data set that is stored in the
distance of the particular hotel from Koyambedu by particular location. R-Tree is built for each location for
making that particular location as the Static user convenient. Figure 3 describes the R-Tree for
location.Table.4. Describes the data set of location dataset in Table.4.
koyambedu. By this similar way N data set of each
location is collected and stored in database.

International Journal of Scientific Research in Science, Engineering and Technology (ijsrset.com)

84
 and their associated texts. In this paper, we use a
style variant of inverted index which is optimized
for third-dimensional points, and is thus named
the spatial inverted index (SI index). This access
technique incorporates point coordinates into a standard
inverted index with little additional space, attributable to
a delicate compact storage theme. Meanwhile, an SI-
index preserves the spatial locality of information points,
and comes with an R-tree designed on each inverted list
at space overhead. Moreover, as a result of the SI-
index depends on the quality technology of inverted
index, which satisfy the user query with quick response
Figure.3: R-Tree for Koyambedu
using key values along with keyword search in the
database.
Figure.3. is the R-Tree for location koyambedu i.e. here
the root of the tree is location which act as a static point III. REFERENCES
location and the hotels in that particular location are the
child of the tree. [1] S. Agrawal, S. Chaudhuri, and G. Das. Dbxplorer: “A system
for keyword based search over relational databases. “In Proc.
Of International Conference on Data Engineering (ICDE),
C. Result pages 5–16, 2002.
[2] N. Beckmann, H. Kriegel, R. Schneider, and B. Seeger. “The
New User registers them by giving their details in the R*tree: An efficient and robust access method for points and
rectangles.” In Proc. of ACM Management of Data (SIGMOD),
user registration page. After registration user gets its pages 322–331, 1990.
own ID and password to access the application to search [3] G. Bhalotia, A. Hulgeri, C. Nakhe, S. Chakrabarti, and S.
the nearest hotel form his location with his required food Sudarshan. “Keyword searching and browsing in databases
using banks.” In Proc. of International Conference on Data
to have. For example, User searches with the associated Engineering (ICDE), pages 431–440, 2002.
text i.e. location as “KOYAMBEDU” and its food as [4] X. Cao, L. Chen, G. Cong, C. S. Jensen, Q. Qu, A. Skovsgaard,
“NORTH INDIAN” in the web page. D. Wu, and M. L. Yiu. “Spatial keyword querying.” In ER,
pages 16–29, 2012.
[5] X. Cao, G. Cong, and C. S. Jensen. “Retrieving top-k prestige-
based relevant spatial web objects.”PVLDB, 3(1):373–384,
2010.
[6] X. Cao, G. Cong, C. S. Jensen, and B. C. Ooi. “Collective
spatial keyword querying.” In Proc. of ACM Management of
Data (SIG- MOD), pages 373–384, 2011.
[7] B. Chazelle, J. Kilian, R. Rubinfeld, and A. Tal. “The bloomier
filter: an efficient data structure for static support lookup
Figure.4: Query Processing Using Key Value
tables.”In Proc. of the Annual ACM-SIAM Symposium on
Discrete Algorithms (SODA), pages 30–39, 2004.
Now it retrieves both the hotels which are near to [8] Y.-Y. Chen, T. Suel, and A. Markowetz. “Efficient query
processing in geographic web search engines.”In Proc. Of ACM
koyambedu bus stand along with user required food Management of Data (SIGMOD), pages 277–288, 2006.
“NORTH INDIAN” at the same time. This is achieved [9] E. Chu, A. Baid, X. Chai, A. Doan, and J. Naughton.
by the key value which is assigned to each location. “Combining keyword search and forms for ad hoc querying of
When the user gives the location as keyword to search databases.” In Proc. of ACM Management of Data (SIGMOD),
2009.
for hotels which in turn call the key value that is [10] G. Cong, C. S. Jensen, and D. Wu. “Efficient retrieval of the
assigned to that particular location and searches the top-k most relevant spatial web objects.”PVLDB, 2(1):337–
database through the key value and retrieve the data that 348, 2009.
[11] C. Faloutsos and S. Christodoulakis. “Signature files: An access
satisfy the user requirement from the data set in the method for documents and its analytical performance
database. evaluation.” ACM Trans- actions on Information Systems
(TOIS), 2(4):267–288, 1984.
[12] I. D. Felipe, V. Hristidis, and N. Rishe. “Keyword search on
II. CONCLUSION spatial databases.” In Proc. of International Conference on Data
Engineering (ICDE), pages 656–665, 2008.
[13] R. Hariharan, B. Hore, C. Li, and S. Mehrotra. “Processing
Now days, Google search engine plays a vital role in spatial keyword (SK) queries in geographic information
case of searching the particular thing that the user want retrieval (GIR) systems.” In Proc. of Scientific and Statistical
Database Management (SSDBM), 2007.
with user query. Conventionally, queries solely focus on [14] D. Zhang, Y.M. Chee, A. Mondal, A.K.H. Tung, and M.
object‟s geometric properties. We have seen some Kitsuregawa, “Keyword Search in Spatial Databases: Towards
fashionable applications that have an ability to pick Searching by Document,” Proc. Int‟l Conf. Data Eng. (ICDE),
pp. 688-699, 2009.
objects supported each of their geometric coordinates

International Journal of Scientific Research in Science, Engineering and Technology (ijsrset.com)

85