Harnessing Machine Learning to Improve the Student Recruitment Process

Session Abstract: As an open enrollment community college, Ivy Tech receives over 150,000 applications every year across 19 campuses. Enrollment teams at each campus have limited recruitment specialists, and they divide and process applications for outreach. With a predictive approach, our goal was to prioritize the applications of new first-time or transfer students in the order of conversion likelihood. We studied student level predictors such as involvement, responsiveness, and demographics and built a model using machine-learning techniques. Our model provided a prediction accuracy of 88%, and the administration has found value thus far even early in deployment. In this presentation, we will share learnings and strategies that we leveraged to address real-world considerations such as iterative model improvements; reacting to changes in upstream/downstream data; data quality and performance monitoring; training enrollment teams to interpret data; and diversity and equity in predictive modeling.

Keywords: recruiting, predictive analytics, machine learning, Salesforce, Snowflake, equity