CLASSIFICATION OF SENTIMENTS USING SOME MACHINE LEARNING METHODS FOR VIETNAMESE TEXT

Abstract

This paper uses several different machine learning methods to evaluate the sentiment classification capability for Vietnamese datasets. The datasets consist of online comments in the field of tourism. Additionally, the experiments compare and evaluate the sentiment classification results of the comments when applying semantic enhancement techniques for Vietnamese texts. The datasets used in the experiments were collected from Facebook fanpages in the field of tourism and online review websites such as Tripadvisor.com.vn and Foody.vn. The experiments use four machine learning algorithms: K-Nearest Neighbor, Support Vector Machines, Naïve Bayes, and Decision Tree. The results showed that the Support Vector Machines machine learning method provides the best sentiment classification performance compared to K-Nearest Neighbor, Naïve Bayes, and Decision Tree methods. This paper is valuable for sentiment classification applications in the field of tourism.