Adaptive Recommendation System in Education Data Mining using Knowledge Discovery for Academic Predictive Analysis: Systematic Literature Review

Abstract

Recent studies of predictive data mining for academics typically produce datasets and methods that allow them to explore data from the education system and using this method is expected to better understand students and the classification in which they study. Many published datasets and methods on predictive data mining for academics are complex and different so that a comprehensive picture of the status of predictive data mining research is lost. Identifying and reviewing trends, topics, data sets, methods and answering research questions regarding predictive data mining for academia between 2017 and 2020 is the aim of this systematic review. Based on data search, 14 related studies were selected to be further identified. The analysis of the selected studies shows that eight topics are the focus and trend topics, namely classification, recommender systems, educational data mining, knowledge discovery databases, student prediction, academic performance, predictive analysis, and predictive analysis. The use of datasets in the selected studies shows that 100% of studies use private datasets and 0% of studies use public datasets. Of the twenty-one methods, there are five methods which are the methods with the most research users in the field of data mining education prediction. Identifies no strong consensus on which algorithm performs best when the study is viewed individually. Therefore, predictive studies on data mining for academics are trying as optimally as possible in choosing an algorithm model to produce more optimal predictions.