랜덤 포레스트를 활용한 영어 내신 성적 관련 예측 요인 탐색

Exploring factors affecting the school-based English exam scores based on Random Forest

Due to the recent change in the admission system, the importance of high school record has increased. The priority on the assessment is the grade and the comment from teachers. Under these circumstances, the purpose of this study was to find factors that were highly correlated with English grades. To this end, the study was designed to analyze factors that predict the English grade of high school record. This study used Korean Education & Employment Panel data and developed a random forest model. In this prediction model, the grades of 4,740 students in the first semester of the third grade were used. The number of variables were 98 and they were classified into three categories (school education, private education, background). Result of the study showed successful performance in other subjects, aiming at achieving post-graduate degrees and the ability of self-directed learning were presented as most highly relevant. Based on the result of this study, teachers should construct the class and the evaluation tools with consideration of the level of student’s self-study. Students should set a specific goal for final education and make learning strategies according to the feature of evaluation system.