New Framework of Educational Data Mining to Predict Student Learning Performance
Educational systems designed to meet the needs of academic advisors about adaptive learning will always be an essential issue, as this will be the beginning of the development of intelligent learning methods. In an educational institution, such as in a university environment, academic guidance carried out by a teacher to his students significantly affects the student's performance in the lecture stage, where educational guidance that goes poorly is allegedly causing difficulties for the student in carrying out his studies, or worst chance of dropping out of school. Therefore, this study aims to explore the potential and capabilities contained in the features of Educational Data Mining to predict students' learning performance which will later present various recommendations for academic guidance methods based on data analysis related to academic records and social and economic related data. In this study, we will propose data analysis and testing from recorded student data in an information technology class from a private university in Jakarta. The modelling presented in this study uses the Decision Tree, Neural Networks, and Naïve Bayes methods, which then implement these algorithms on academic data from 300 students of the 2017-2019 and 2018-2020 Information Systems and Informatics study program. From the implementation of data mining techniques in this study, performance results were obtained, which stated that the designed framework provided accurate predictions related to student performance.