Step 2. Use machine learning to identify ‘at risk’ or disengaged students

The next step is to extract the data required to build and continually retrain a machine learning model. OES’s predictive modelling leverages students’ behavioural, academic and demographic data: engagement metrics collected daily, performance metrics collected per assessment and student characteristics collected during admission. “The intention is to identify and reach out to students with meaningful support, before they fail, defer or drop out.”