The Comparison of Data Mining Algorithms for Classification of Suggestions from Computer-Based Written Exam Participants

摘要

The Computer-Based Written Exam is a mandatory requirement for taking the State University Joint Entrance Test. As a form of implementation evaluation, each test participant is required to comment on the questionnaire given. The number of comments collected was 663 comments. Data processing used data mining methods. One form of utilization of data mining is classification. These comments were classified into 10 categories. This study compared the performance of classification algorithms based on parameter values of accuracy, precision, and recall. The classification algorithm compared was Naïve Bayes, Support Vector Machine (SVM), and Decision Tree. The SVM algorithm had better performance with an average value of accuracy, precision, and recall of 77.46%, 77.05%, and 74.65%.