A Survey on Sentiment Analysis using Swarm Intelligence

The social web data has increased tremendously in the recent years in form of comments, blogs, reviews and tweets. The nature of this data is highly un-structured and high- dimensional, making text classification a tedious task. Sentiment analysis, which is a text classification technique is applied on this data to gauge user opinion on several pertinent issues. As a natural language processing task, sentiment analysis automatically mines attitudes or views of users on specific issues. It is a multi-step process where selecting and extracting features is a vital step that controls performance of sentiment classifier. The statistical techniques of feature selection like document frequency thresholding produce sub optimal feature subset due to the Non Polynomial (NP) hard nature of the problem. Swarm intelligence algorithms are extensively used in optimization problems. Optimization techniques could be applied to feature selection problem to produce Optimum feature set. Swarm Intelligence algorithms are used in feature subset selection for reducing feature subset dimensionality and computational complexity thereby increasing the classification accuracy. In this paper we study the state-of-art of the various swarm intelligence algorithms which are presently used for feature subset selection within the sentiment analysis framework. The study shows that swarm optimization brings significant accuracy gains. There are only few swarm algorithms which have been applied in this area and there are many other algorithms which can be explored, this study provides an insight into the various algorithms which can be expounded for improved sentiment analysis.