A Machine Learning Based Approach for Opinion Mining on Social Network Data

Abstract

Micro-blogging has been widely used for voicing out opinions in the public domain. One such website, Twitter is a point of attraction for researchers in the areas such as prediction of electoral events, movie box office, stock market, consumer brands etc. In our paper, we focus on using Twitter, for the task of opinion mining. We explore how combining the different parameters affect the accuracy of the machine-learning algorithms with respect to the consumer products. In this paper, we have combined the methods of feature extraction with a parameter known as negation handling. Negation words can awfully change the meaning of a sentence and hence the sentiment expressed in them. We experimented with supervised learning methods like Naïve Bayes (NB) Classifier and Maximum Entropy (MaxEnt) Classifier along with optimization iteration algorithms i.e., Generalized Iterative Scaling (GIS) and Improved Iterative Scaling (IIS). Experimental evaluations show that our proposed technique is better. We have obtained a 99.29% of specificity measure using the MaxEnt-IIS Classifier.