IMPROVING PERFORMANCE OF STUDENTS’ GRADE CLASSIFICATION MODEL USES NAÏVE BAYES GAUSSIAN TUNING MODEL AND FEATURE SELECTION

Abstract

Student grades are a relevant variable for predicting student academic performance. In achieving good and quality student performance, it is necessary to analyze or evaluate the factors that influence student performance. When a educator can predict students' academic performance from the start, the educator can adjust the way of learning so that learning can run effectively. The purpose of this research is to study how it is applied to determine the interrelationships between variables and find out which variables have an effect, then use it as a feature selection technique. Then, researchers review the most popular classifier, Gaussian Naïve Bayes (GNB). Next, we survey the feature selection models and discuss the feature selection approach. In this study, researchers will classify student grades based on existing features to evaluate student performance, so it can guide educators in selecting learning methods and assist students in planning the learning process. The result is that applying Gaussian Naïve Bayes (GNB) without feature selection has a lower accuracy of 10.12% while using feature selection the accuracy increases to 10.12%.