Supervised Machine Learning for Prediction of Minimum Completeness Criteria (KKM) Scores for Elementary School Students

Mustakim
Rahim, Arham

Jurnal Penelitian Pendidikan IPA (JPPIPA)
Vol 10 No 11 (2024): November

This study aims to predict potential declines in students' Minimum Completeness Criteria (KKM) in higher grades (4th, 5th, and 6th) by analyzing their cognitive, affective, and psychomotor scores from lower grades (1st, 2nd, and 3rd). Using a quantitative research method, various machine learning algorithms were applied, including Naive Bayes, K-Nearest Neighbors (KNN), Support Vector Machines (SVM), and Neural Networks. The dataset comprised students' scores across cognitive, affective, and psychomotor domains from the lower grades. After training and comparing the models, the Neural Network algorithm demonstrated the best performance, achieving 89% accuracy and 100% recall. These results indicate that the model can help teachers identify students at risk of struggling with KKM standards in higher grades, enabling early interventions. The study concludes that Neural Networks offer a promising tool for early detection of academic challenges in elementary education.