Hybrid Sentiment Analyzer for Opinion Mining Indian Admission Scenario

Abstract

Social media has become one of the widely acclaimed tool for sharing information as well as expressing ideas and emotions. The work depicts the dual aspect task of analyzing and comprehending data available on Twitter platform. This is done using NLP techniques. Using Latent Dirichlet Allocation (LDA) topic technique; the major topics discussed in tweets (of data set taken), have been identified. The input for this Latent Dirichlet Allocation is given by NLP technique – Bag of Words. For further processing, identification of the underlying emotions contained in tweets using the techniques of Sentiment Analysis is done. The result of sentiment analysis is in the polar form. As a case study, a scenario of admissions in India for UG and PG has been considered. The whole process has captured the opinions of stake holders taking part in the admission process. Tweeter data of Indian Institute of Technology (IIT) admission has been used to collect the data in order to conduct the experiment. Major topics discussed in tweets and the fundamental emotions contained are obtained as results along with the polarity of the tweets.