Advanced Twitter Sentiment Analysis Using Supervised Techniques and Minimalistic Features

Abstract

Social media along with different web forums contribute to a huge amount of data, i.e. opinions, feedbacks, reviews generated on a daily basis. Sentiment analysis is the identification of polarity (positive, negative or neutral) of the data to analyse the opinion of users on various platforms on a range of topics including those that have real business value, i.e. polarity of users on a product. With the advent of platforms like Twitter, people freely express their opinion on almost everything and that generates a huge amount of data which cannot be processed or analysed manually; therefore, we have various NLP and machine learning techniques which can effectively analyse and predict the polarity of the data which enables us to capture the sentiment of the people regarding a particular issue. In this paper, we intend to analyse the different machine learning algorithms for sentiment analysis of Twitter data and compare the algorithms’ performance on different datasets with the help of metrics like precision, accuracy, F-measure and recall.