Comparison Of The Naïve Bayes Algorithm And The Decision Tree On Sentiment Analysis Of Student Comment Data In The Digital Teacher Assessment (DITA) Application

Abstract

Telkom Purwokerto Vocational High School (SMK) is a school managed by the Telkom Education Foundation. They use Digital Teacher Assessment (DITA) apps. Students provide comments to the teacher using the DITA application. The collected comment data will be grouped into three categories, namely positive, negative, and neutral. Based on the category of comments requires sentiment analysis in grouping these comments. Sentiment analysis uses lexicon-based. After getting sentiment analysis using lexicon-based, then the words are weighted using TF-IDF and then classified and evaluated. This study uses an algorithm naïve Bayes and a decision tree. So the results of the comparative research on the accuracy of the naïve Bayes algorithm and the decision tree with the decision tree algorithm have the highest level of accuracy, namely 99%. So it can be concluded that using the decision tree algorithm is better at classifying student comment sentiment analysis data.