Review of Effective Models for Sentiment Analysis

The sentiment analysis of data has become rather important in today's world where people share their opinions and views freely on social media platforms such as Twitter. Companies use various sentiment analysis models to analyse the data available on social media. However some of these models are unable to discern the contextual meanings of words in a sentence and can end up classifying erroneously. This paper looks to suggest ways to improve the sentiment analysis models and their accuracy of classification of tweets as negative or positive. One way is by combining different single models to overcome shortcomings of a single model [1]. The other is classification on the basis of features such as negation and conjunctions [2].