Sentimental Analysis of Online Reviews Using Fuzzy Sets and Rough Sets

Sentimental Analysis of Online Reviews Using Fuzzy Sets and Rough Sets

Anuradha Jagadeesan (VIT University, India) and Amit Patil (VIT University, India)
DOI: 10.4018/978-1-5225-1008-6.ch017
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

With the increased interest of online users in E-commerce, the web has become an excellent source for buying and selling of products online. Customer reviews on the web help potential customers to make purchase decisions, and for manufacturers to incorporate improvements in their product or develop new marketing strategies. The increase in customer reviews of a product influence the popularity and the sale rate of the product. This lead to a very important question about the analysis of the sentiments (opinions) expressed in the reviews. As such internet does not have any quality control over customer reviews and it could vary in terms of its quality. Also the trustworthiness of the online reviews is debatable. Sentiment Analysis (SA) or Opinion Mining is the computational analysis of opinions, sentiments, emotions and subjectivity of text. In this chapter, we take a look at the various research challenges and a new dimension involved in sentiment analysis using fuzzy sets and rough sets.
Chapter Preview
Top

Introduction

Today shopping online has become more popular than ever. Online shoppers use Customer’s reviews to make informed buying decisions. Consumer actively seeks out and reads customer reviews prior to making a purchase decision. According to “social shopping study” by Power Reviews 70% of online shoppers, accepted customer reviews and ratings on a retailer’s website were extremely important when they are selecting or purchasing product.

Manufactures use these reviews to know consumer preferences and interests, to maximize the profit. It is important to know others thoughts before we make purchase decision. As online purchasing increases, number of reviews the web site receives also increases. Hence, the number of customer reviews being posted at review site is growing at faster rates (in hundreds or thousands). Maintaining such huge amount of reviews is not only a problem for the manufacturers but makes the process of decision making complicated, causing customer either to read all the reviews on that product or to select reviews and read. It is time consuming to read all the reviews and to select the reviews and read; the customer does not know which reviews to choose and read since the quality of the reviews vary greatly. Figure 1 shows the summary of Social shopping study by power reviews.

Figure 1.

Social study chart on customer feedback

As text messages express the state of minds of individual; mining such messages from a large population in different context. Discovering user preferences through social media is a challenging task. Opinion mining, SA and subjectivity analysis are related fields sharing common goals of developing and applying computational techniques to process collections of opinionated texts or reviews. Opinion mining (OM) is a new field of data mining concerned with the opinion that can be induced from documents. OM is divided into three major tasks: development of linguistic resources, sentiment classification, and opinion summarization (Elawady, Barakat, & Elrashidy,2014).

Sentiment Analysis (SA) is a task that finds the opinion (e.g .positive or negative) from the text document like product reviews/movie reviews. As user generated data is increasing day by day on the web. It is needed to analyze those contents to know the opinion of the users, and hence it increases the demand of sentiment analysis research. People express their opinions about movies and products etc. on the web blogs, social networking websites, content sharing sites and discussion forums etc. Companies can use sentiment analysis in improving their products based on the user's feedback written about their product on blogs, E-commerce based companies know the online trends about the product. For example, sentiment analysis knows which model of camera is liked by the most users. Sentiment classification can also be considered as a text classification problem. Bag of Words (BOW) representation is commonly used for sentiment classification using machine learning approaches. The words present in all documents create feature vector. Generally, this feature vector is a huge in dimension that is used by machine learning methods. The concepts of fuzzy sets and rough sets could be used as classification techniques to improve the polarity of expressed sentiments.

Top

What Is Sentiment Analysis?

Sentiment Analysis is a Natural Language Processing and Information Extraction task that aims to obtain writer’s feelings, emotions expressed in positive or negative comments, questions and requests, by analyzing a large numbers of documents. Speaking in general terms, sentiment analysis aims to determine the attitude of a speaker or a writer with respect to some topic or the overall polarity of a document. With the increase in online users and their exchange of public opinions is the driving force behind Sentiment Analysis. The Web is a huge repository of structured and unstructured data which varies in its quality. The analysis of this data to extract opinion (sentiments) and classify them as positive, negative or neutral is a challenging task.

The work of Liu (2010) in NLP handbook defines a sentiment or opinion as a quintuple(oj, fjk, soijkl, hi, tl),where

Complete Chapter List

Search this Book:
Reset