A Hybrid Approach for Enhancing Accuracy and Detecting Sarcasm in Sentiment Analysis

Sentiment analysis and opinion mining are gaining significant importance in the field of research. In last few years, many efforts had been made to mine the opinions and sentiments of people which they post on internet in the form of reviews, blogs, forum, discussions and micro-blogs. The important thing is to discover what people actually think. Their opinions are classified on the basis of polarity which can be positive, neutral or negative. In this study we have given a brief overview of sentiment analysis and a hybrid approach which includes different classifiers such as emoticon classification, negation and Modifier classification, Hashtag classification, Sarcasm Detection using deepmoji which is an advanced Sentiment Analysis technology, Polarity detection using SentiWordNet and Domain Specific Classifier.