Systematic Review and Analysis of EDM for Predicting the Academic Performance of Students

In education sector, two main tasks are—Analysis of students’ data and predicting their performance as it may leads to their great future. Analyzing the data obtained from educational institute or Learning Management System (LMS) becomes challenging task among instructors to predict the performance of students correctly and accurately. This article present the review of 231 research articles which suggest the various methodologies like classification, clustering, association rule and other methodologies like Fuzzy System, Big Data, Text Mining, Artificial Intelligence, Sentiment Analysis, Pattern Mining, Process Mining,Graph Mining, etc. based on students’ performance. The research articles are reviewed on the basis of various Data Mining techniques such as supervised learning—classification algorithms, unsupervised learning—clustering and association rule mining and other techniques also. The research gap and issues faced by existing techniques in Educational Data Mining (EDM) are described and deliberated which researcher working in this area can use it for future works. The work are analyzed based on purpose in EDM, Technique used in EDM, Classification Techniques, Classification performance metric, selecting the best Classification technique, Clustering Techniques, Association Rule Techniques, parameter metrics related to Association Rule Mining, open source/ commercial software used to develop model in education sector, and data mining tools. The commonly used technique for the prediction of academic performance of students in education stream are classification technique—Naïve Bays, J48, and Support Vector Machine, clustering technique—K-means clustering algorithm and the association rule mining algorithm—Apriori association rule mining algorithm. Inspire of lot of research in EDM, still there are some limitations related to the prediction of students’ performance. So this article will be helpful to the researchers for students’ academic performance prediction based on the data retrieved from e-learning, LMS, etc. correctly and accurately.