MONITORING LEARNERS’ PERFORMANCE BY MODELING LEARNING PROGRESS USING MACHINE LEARNING

In this globalization era, the involvement of machines is inevitable, including one within the context of monitoring learners’ performance. This study investigates learners’ performance monitoring to predict learning progress using Machine Learning techniques. This study uses the results of a one-semester assessment of Indonesian junior high school students in mathematics. Aspects of the assessment include cognitive and psychomotor domains while the forms of assessment comprise formative and summative. This study shows that learning progress modeling can also be carried out throughout learning and model accuracy is fairly good in four learning periods. The learning progress model generated in the last period of the course is better than that in the early period of the course. Out of 7 machine learning algorithms being compared, Random Forest shows the best accuracy as it reaches 92% in the whole period. It is indicated that the model being used in this study is proven to be effective in one subject being taught in one semester.