Pengembangan Sistem E-Learning Berbasis Decision Tree untuk Analitik dan Klasifikasi Nilai Siswa di SDK Harapan Denpasar

Abstract

Monitoring student academic performance at the elementary school level remains a challenge because most grade management is still performed manually, slowing the identification of students who need special attention. This study developed a Laravel 10-based e-learning system integrating online examination with security features, automatic grade recapitulation, and academic performance classification using the CART Decision Tree algorithm. The system was built using the waterfall model and evaluated on 60 students from grades 1A to 6A at SDK Harapan Denpasar. Final scores were computed using a 40% assignment and 60% examination weighting formula, then classified into four categories: Very Good, Good, Sufficient, and Needs Guidance. Model evaluation yielded an accuracy of 86.7%, macro precision of 89.6%, macro recall of 86.3%, and macro F1-score of 86.2%, all exceeding the 80% threshold. Black box testing passed 100% of test cases (15/15) and User Acceptance Testing produced a score of 4.68 out of 5.00. The system enables teachers to identify students requiring further guidance directly from the dashboard without relying on external analytical tools.