Systematic Mapping Study of Literature on Educational Data Mining to Determine Factors That Affect School Performance

This document presents the results of conducting a Systematic Mapping Study (SMS) on the mining of educational data to determine factors that affect school performance in the higher education system. As a result of this analysis, 20 primary studies were obtained, where it is observed how data mining is applied to predict school performance and thus decrease the dropout rates of students. Through this work a more integral perspective is offered from the process of information capture and automatic consolidation until proposing a prediction model based on the analysis and use of data mining techniques, contributing with new analytical approaches to the current studies, taking consider the characteristics of educational management. In conclusion, the studies consider that, by using educational data mining applied to large volumes of data existing in higher education institutions, it would improve decision making to guarantee a high-quality education, highlighting the performance problems from the beginning and proposing corrective actions. In addition, these studies show data mining techniques used, such as logistic regression, decision trees, random forests, Naive Bayes to explore data from educational environments.