Improved sentiment analysis for teaching evaluation using feature selection and voting ensemble learning integration

Teaching evaluation system is widely used to assess and investigate the education quality. Presently, sentiment analysis contributes for student sentiment polarity detection in teaching evaluation which collects the feedback messages. Text mining techniques are broadly extended to classify the effective improvement of the sentiment polarity analysis. Furthermore, the feedback messages from opened-end questions which stored in teaching evaluation system are selected for the classification. In addition, various methods used for classification in the experiment are Naïve Bayes, ID3, J48 Decision tree. In this paper, reducing the feature in data preprocessing stage and teaching sentiment analysis using voting ensemble method of machine learning are proposed and compared with existing typical machine learning for sentiment analysis. The experimental results show that the voting ensemble learning integrate with Chi-Square feature selection exhibits higher than typical classifiers.