Pemodelan Klasifikasi Siswa Berprestasi dengan Random Forest: Studi Kasus pada Bimbingan Belajar

Abstract

Academic achievement is a goal desired by every student, leading many to attend additional lessons at tutoring institutions to improve their learning outcomes. This study aims to classify student achievements at a tutoring institution based on periodic evaluation results using the Random Forest algorithm. The dataset used includes 112 students from the 2017 to 2018 academic year, with 67 student records for training and 45 for testing. Evaluation results indicate that students classified as underachieving dominate (98 students), while only 14 students meet the criteria for high achievement. The analysis shows the highest average scores in English (85.38) and Mathematics (83.66), while the lowest averages are in Social Studies (70.47) and Science (78.96). Applying the Random Forest algorithm to the test data resulted in four students with a maximum confidence score of 0.933, demonstrating that the model has high accuracy and can be utilized by the institution to monitor and motivate students to achieve high-performance categories. This research contributes to the development of data-driven systems to support decision-making processes in tutoring institutions.