A Framework for Sentiment Analysis Classification based on Comparative Study

Abstract: A number of Feature Selection and Ensemble Methods for Sentiment Analysis Classification had been introduced in many searches. This paper presents A frame work for sentiment analysis classification based on comparative study on different classification algorithms i.e., comparison between combinations of classification algorithms: Bayes, SVM, Decision Tree. We also examined the effect of using feature selection methods (statistical, wrapper, or embedded), ensemble methods (Bagging, Boosting, Stacking, or Vote), tuning parameters of methods (SVMAttributeEval, Stacking), and the effect of merging feature subsets selected by embedded method on the classification accuracy. Particularly, the results showed that accuracy depends on the feature selection method, ensemble methods, number of selected features, type of classifier, and tuning parameters of the algorithms used. A high accuracy of up to 99.85% was achieved by merging features of two embedded methods when using stacking ensemble method. Also, a high accuracy of 99.5% was achieved by tuning parameters in stacking method, and it reached 99.95% and 100% by tuning parameters in SVMAttributeEval method using statistical and machine learning approaches, respectively. Furthermore, tuning algorithms' parameters reduced the time needed to select feature subsets. Thus, these combinations of algorithms can be followed as a frame work for sentiment analysis.