A Novel Approach of Sentiment Detection on Twitter

Abstract

Twitter has emerged as a platform to express the opinion on various issues. Plenty of approaches like machine learning, information retrieval and Natural Language Processing have been exercised to figure out the sentiment of the tweets. Each of these methods has some benefits and limitations based on the data type used and suitability of data. Most of the research work has been carried out on application of machine learning algorithms applicable for social media sites and further getting the accuracy of the result. However the machine learning algorithms can be integrated with natural language processing algorithms for refining accuracy and context. These refinements tend to increase the accuracy of the result. In the present thesis, we have purposefully integrated the naive bayes and adjective analysis for finding the polarity of the ambiguous tweets. Experimental outputs have revealed that the overall accuracy of the process is improved using proposed model. Firstly we have applied naive bayes on collected tweets which results in set of truly polarized and falsely polarized tweets. False polarized set has further processed with adjective analysis to determine the polarity of tweets and classify it to be positive or negative. For adjective analysis we have made corpus of adjective negative and positive polarity. We have used movie reviews for our training set as well as test set.