Multi-Class Sentiment Analysis Comparison Using Support Vector Machine (SVM) and BAGGING Technique-An Ensemble Method

Multi-class analysis, as the term suggest is the classification of the data in more than two classes. However not much studies were focused on such analysis and researchers often confined themselves to the binary sentiment classifiers. In this paper, we proposed machine learning algorithm as an approach to predict the sentiment classification. The experiments are conducted on public data sets combined with ensemble method named BAGGING, an abbreviation for Bootstrap aggregation with 10-cross fold validation technique is used to obtain the classification accuracy. The result accuracy suggested the exploring further improvement using the combination of the multi-class sentiment classifiers.