A Machine Learning Framework to Predict College Readiness Using Early Dropout

• In Event: Marriott Roundtable Session Five
  In Roundtable Session: Higher Education: Creating New Paths for Data Use (Table 4)

Abstract

Although states, school districts, and research institutions have historically put in a huge amount of time and resources to develop standards for college readiness and create intervention programs to help students navigate college, college readiness remains a big problem. This study argues for the benefits of using early college dropout to indicate the lack of college readiness and designs a machine learning framework around predicting students at risk of dropping out of college. This study explores features useful for the prediction task, applying and tuning the popular learning algorithms, and evaluating their performance using meaningful metrics. This study discusses the implications of the findings and informs the design of an early warning system and its intended applications.

Author

• Yiwang Li, University of California - Riverside
  Presenting Author