A Comparative Study on Vectorization and Classification Techniques in Sentiment Analysis to Classify Student-Lecturer Comments
Ochilbek Rakhmanov
Abstract
Sentiment analysis is one of the important fields in educational data mining. In this paper, a large dataset, more than 52 000 comments, was used during experiment to develop a state-of-art classification model. The correlation test was conducted on sentiment analysis results and scale-rated survey results, and the result (r(203)=.79, p<.001) shows that sentiment analysis can be accepted as reasonable method for course and lecturer evaluation. A comparative analysis was done between different vectorization and classification techniques. The results of the experiment show that classifier built using Random Forest was most optimal and efficient classification model with state-of-art prediction accuracy of 97% for 3-class classification. Moreover, to improve the diversity of the comments, a 5-class dataset was formed and experiment resulted with an efficient classification model with accuracy of 92%. The Tf-Idf vectorization technique performed better than Count (Binary) vectorization.