Sentiment Analysis of Book Reviews Using Semantic Network

Abstract

In new emerging technologies the processing of large textual data is challenging. Text mining can be defined
as the process where exploring and analysing of enormous amounts of unstructured text data assisted by software that can
identify patterns, concepts, keywords, topics and other attributes in handling the data. From the data, the main challenge
is to identify the meaning of the text and correlate it to the text representation model for the sentiment classification. It is
complex to capture the sentiments in the document level sentences to correctly classify them into positive, negative and
neutral. In this work, we suggest a novel semantic network method using WuPlamer word similarity method of WordNet
to compare the Part-Of-Speech (POS) tagged words of new documents to the manually obtained POS tagged words from
the documents. This results in generating a similarity score, which is then used to classify the documents into three
categories. We experimented to evaluate the different existing algorithms. The results show that the proposed approach
allows for the precision of the classification when considering analysis of the document level.