Predicting Potential School Failure

Providing academic support to students so that they do not fail in school is a critical challenge faced by teachers. It is difficult for teachers to keep track of and integrate the various activities done by students and predict whether a student will pass or fail.

With learning management systems and other digital learning tools, data regarding students’ activities is captured. Machine learning methods applied to this data can then provide insights into student behavior and performance that is otherwise unavailable or not evident to teachers. These algorithms can reliably predict students’ failure in school based on their past and current behaviours and performance. Such a prediction can be leveraged by a teacher to provide extra support to a student who needs it. The goal of this project is to use available student data to build prediction models of student performance based on various behavioural indicators.