Student Performance Prediction in Sebelas Maret University Based on the Random Forest Algorithm

Students who have low levels of academic performance may result in such students having drop out. Various factors influence the level of academic performance of such students. Preventive action would be better to cope with the drop out. This study aims to conduct prediction of students' academic performance at Sebelas Maret University based on three categories of factors namely social, economic, and academic factors. Methods used include, data acquisition stages, data preprocessing, feature selection, classification, and analysis of results. Feature selection uses information gain (IG) and Random Forest (RF) classification algorithms, with 10-fold cross validation. The test results showed an accuracy of 90.7%, such performance outperforms support vector machine (SVM) and Decision Tree (DT) algorithms.