2019 International Conference on Machine Learning, Big Data, Cloud and Parallel Computing (Com-IT-Con), India, 14th -16th

Feb 2019

Sentiment Analysis on Customer Feedback Data:

Amazon Product Reviews
Pankaj, Prashant Pandey, Muskan, Nitasha Soni
Manav Rachna International Institute of Research and Studies,
Faridabad, Haryana
[email protected], [email protected], [email protected],
[email protected]

Abstract- Sentiment analysis is one of the fastest spreading star scaled system, where the highest rating has 5stars and the
research areas in computer science, making it challenging to lowest rating has only 1star Figure1.
keep track of all the activities in the area. We present a customer
feedback reviews on product, where we utilize opinion mining,
text mining and sentiments, which has affected the surrounded
world by changing their opinion on a specific product. Data used
in this study are online product reviews collected from
Amazon.com. We performed a comparative sentiment analysis of
retrieved reviews. This research paper provides you with
sentimental analysis of various smart phone opinions on smart
phones dividing them Positive, Negative and Neutral Behaviour.

Keywords: sentiment analysis, text mining, opinion mining, product

reviews.
I. INTRODUCTION
Opinions are statements that reflect people’s perception or Here, the report tackles fundamental clauses of sentiment
sentiment. Sentiment analysis is a series of methods, analysis, namely sentiment polarity distribution. This paper
techniques, and tools about detecting and extracting subjective prospective a schema for online opinion representation. The
information, such as opinion and attitudes, from language [1], inputs to the schema are product name, date and time and
helping in finding the mood of the customers about a review of that product, where output contains the summary of
purchasing of a particular product or topic. It involves the review in compact manner. The whole process includes
building a system to collect and examine opinions about the summarization in three steps: (1) Product feature based, which
product made in many online purchasing sites. Sentiment is given by customer; (2) In each review, Identify expected
analysis is a sub field of web content mining. This paper is features in each opinion sentence and (3) Finding out whether
organized as follows: Section I Introduction, Section II the feature/opinion is positive, negative or neutral and finally
Related Work, Section III Opinion mining and Sentimental summary will be created [4]. This paper, propose a general
analysis, Section IV Sentimental Classification, Section V method for finding opinion features from online reviews by
Opinion Mining Techniques, and Section VI Tools Used in exploiting the difference in opinion feature statistics across
opinion mining, Section VII Applications, Section VIII two compilation, an domain-specific corpus and one domain-
Research challenges and in the last Section Research Scope. independent corpus[5]. This product review paper discusses
existing techniques and approaches for feature extraction in
II. RELATED WORK sentimental analysis and opinion mining [6]. This paper gives
the feature based opinion mining and their efficiency in terms
Conventionally, sentiment analysis has been about opinion of precision, recall and accuracy. In this paper first extracts
contradiction, i.e., whether someone has positive, neutral, or the feature, modifier and opinion.
negative opinion towards something [2]. Data used in this
paper is a combination of product reviews collected from III. OPINION MINING AND SENTIMENT
Amazon.com [3], between July and September, 2018. There ANALYSIS
have been somewhat overcome in the following two manners:
Firstly, each product review holds inspections before it can be Sentiment analysis, also known as Opinion mining, is the
posted. Secondly, respective review must have a rating on it study of sentiments that determines the judgement of people’s
that can be used as the ground truth. The rating is based on a opinions, sentiments, evaluations, and emotions in relation to
entities such as products, services, organizations, events,
topics and their different attributes. Generally opinion cannot

978-1-7281-0211-5/19/$31.00 2019 ©IEEE 320

structure a problem but it can subjective and in case opinion
gathered from many people it should be summarized. Product categories
The approach of an opinion mining is described by
[Jin.2006, Liu, 2010]. They put most impact on their 5
work and finds the basic components of an opinion are: 4
Opinion holder: it is the person who gives a specific 3
opinion on a particular object.
2
Object: it is called as the entity over which an opinion is
expressed by user. 1
Opinion: it is a view, sentiment, or assessment of an 0
object done by user. Mobiles Computers Flashdrives Electronics
Fig. 2. Product Categories
Opinions are of two types described as: Regular and
Comparative. Regular opinion is expressions on some B. Sentiments Sentences Judgement and POS Tagging
target entities, which can be further classified into direct This process is proposed by Pang and Lee [8] in which all
and indirect opinion. Comparative opinions are the objective content should be removed for analysis of sentiment.
Comparisons of more than one entity [7]. Instead of removing objective content, in our study, all
subjective content was extracted for future analysis which
consists of all sentiment sentences. A sentiment sentence is
that which contains, at least, one positive or negative word.
All of the sentences were first of all arranged into categorized
English words. Every word of a sentence has its semantic role
that defines how the word is used. The semantic roles are also
called as the parts of speech. There are generally 8 parts of
speech in English: the verb, the pronoun, the noun, some
adverbs and prepositions, the interjection, and the conjunction.
In natural language preference and suitability, part-of-speech
(POS) taggers [9-10] have been generated to classify words
Fig. 1. Opinion Mining Work Flow based on their parts of speech. In sentiment classification, a
POS tagger is most important because of the following two
1. Direct opinions: This includes Text documents that give reasons: 1) Words like nouns and pronouns mostly do not
positive or negative opinion about the product directly. contain any sentiment. So, it is able to filter out such words
Example: “Battery backup of this mobile is too bad. with the help of a POS tagger; 2) A POS tagger can also be
2. Comparison Opinions: Opinions found in text document useful in distinguish words that can be used in different parts
and are meant to compare the object with some other of speech. For instance, as a verb, “improved” may conduct
objects. Opinion mining is the part of the web mining. different amount of sentiment as being of an adjective. The
POS tagger used for this survey is a max-entropy POS tagger
IV. RESEARCH PATTERN AND METHODOLOGY developed for the Penn Treebank Project . This tagger is able
to provide 46 different tags which indicate that it can identify
A. Collection of Data more detailed semantic roles than only 8.
The appropriate amount of Data used in this paper is a
arranged set of product reviews collected from V. CONCLUSION
amazon.com. From August to December 2018, in total,
we collected over 500 sentiments of product reviews in Sentiment Analysis or opinion mining is a case study which
which the products belong to 4 major categories: Mobiles, analyses people’s sentiments, attitudes, entropy or emotions
Computers, Flash drives and Electronics 3(a)). These towards certain entities. This paper tackles a fundamental
online reviews were posted by over 3.2 millions of problem of sentiment analysis, sentiment polarity
customers (reviewers) towards 10,001 products. Each categorization. The data for this research is collected of online
review includes the following information: 1) product reviews from Amazon.com. A process known as
reviewer_ID; 2) product model; 3) date and time of the sentiment polarity categorization and POS has been proposed
review; 4) review text. along with detailed descriptions of each step. These steps
consist of pre-processing, pre-filtering, biasing, data accuracy
etc. features which require the knowledge of machine
learning. A lot of work in opinion mining and sentiments of
customer reviews has been conducted to mine opinions in
form of document, sentence and feature level sentiment
analysis. For future preferences, Opinion Mining can be

321
carried out on set of discovered feature expressions extracted
from reviews become a most interesting research area. There
is more innovative and effective techniques have to be
invented which should overcome the current challenges faced
by Opinion Mining and Sentiment Analysis.

REFERENCES
[1] B. Liu, Handbook Chapter: Sentiment Analysis and Subjectivity,
Handbook of Natural Language Processing, Handbook of Natural
Language Processing, MarcelDekker, Inc.NewYork, NY,
USA,2009
[2] K. Dave, S. Lawrence, and D. M. Pennock, “Mining the peanut
gallery: Opinion extraction and semantic classification of product
reviews,” in Proceedings of the 12th international conference on
World Wide Web, 2003, pp. 519–528.
[3] Mugdha More and Bharat Tidke, July 2015. “A Framework For
Summarization Of Online Opinion Using Weighting Scheme”,
ACII, Vol.2, No.3.
[4] Haseena Rahmath P,” Opinion Mining and Sentiment Analysis
Challenges and Applications”, IJAIEM Volume 3, Issue 5, May
2014.
[5] Muhammaad Zubair, Aurangzeb Khan, Shakeel Ahmad, Fazal
Masud Kundi and Asghar, 2014. ”A Review of Feature
Extraction in Sentiment Analysis”. ISSN 2090-4304 Journal of
Basic and Applied Scientific Research.
[6] Buche A., Chandak M.B., Zadgaonkar A.(2013).Opinion mining
and analysis: a survey. International Journal on Natural Language
Computing (IJNLC).39-48.
[7] Pang B, LeeL (2004) A sentimental education: Sentiment analysis
Using subjectivity summarization based on minimum cuts. In:
Proceedings of the 42 Nd Annual Meeting on Association for
Computational Linguistics, ACL’04. Association for
Computational Linguistics, Stroudsburg, PA, USA
[8] RothD, ZelenkoD (1998) Part of speech tagging us in ganetwork
of linear separators. In: Coling-Acl, The 17th International
Conference on Computational Linguistics.pp1136–1142
[9] KristinaT (2003)Stanfordlog-linearpart-of- Speech tagger.
http://nlp.stanford.edu/software/tagger.shtml
[10] MarcusM (1996) Upenn part of speech tagger.
http://www.cis.upenn.edu/~treebank/home.html

322