Application of Data Mining to Predict Student Learning Outcomes in Padang Panjang

According to the 2022 Annual Best Countries survey, Indonesia's educational quality lags behind neighboring countries like Singapore and Malaysia. Students' difficulty meeting minimum grade requirements is a significant issue which can lead to expulsion. Educational institutions collect vast amounts of data annually, including student demographics and semester grades used for evaluating academic performance. However, this data is often underutilized. This study aims to predict student academic performance based on sociodemographic data and average semester grades. The study focuses on 643 students from a vocational high school in Indonesia. Ten attributes with student grades are used as the label to determine performance. The Decision Tree C4.5 and Naïve Bayes analysis using RapidMiner. The accuracy using C4.5 and Naïve Bayes was obtained at 78.12% and 76.88%, respectively. Notably, “Gender” was identified as the most influential factor in student performance at the school. The generated rules can assist schools in taking preventive measures to reduce the number of students struggling to meet minimum grade requirements.