期刊名称：International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
印刷版ISSN：2278-1323
出版年度：2017
卷号：6
期号：8
页码：1234-1239
出版社：Shri Pannalal Research Institute of Technolgy
摘要：Nowadays, most of the people generally used to give their opinions or their views on social sites like facebook, twitter, etc. Out of these, Twitter is mostly used social site which is practical for analyzing the sentiments about any query term. In this research work, the query is about Punjab elections 2017 regarding which we have taken a corpus of tweets about the query mentioned. Corpus are used by two or three annotators for the annotation process to determine its polarity about the three major parties in Punjab elections, 2017 i.e. Aam Aadmi Party (AAP), Congress party and SAD-BJP party. The positive or negative sentiments of tweets show that people either are in favor of the party or against it, if neither it is positive nor negative then it is considered as a neutral. After annotation, there are four classifiers we used i.e Lexicon approach, Naïve Bayes, Naïve Bayes with Binary features and Extension of Naïve Bayes using lexicon which define the positive, negative and neutral sentiments. After this, standard measure of evaluation (precision, recall, and f score) are used to find the accuracy of every approach.
