Use Educational Data Mining to Predict Undergraduate Retention

Abstract

This paper presents an application of educational data mining to predict undergraduate retention. The research provides valuable insight about data feature ranking, algorithm selections and validation methods based on unique types of data that come from educational settings. The data from a cohort of 972 students enrolled in 2008 at Embry-Riddle Aeronautical University (ERAU) were used to train and validate the predictive models. This research aims to provide decision recommendations to ERAU and similar institutions to make the timely intervention for improving retention.