Predicting Joining Behavior of Freshmen Students using Machine Learning – A Case Study

With the increasing competition, universities are trying to reach out to aspiring students to get them enrolled. However, out of all the students enrolled to a university, many do not actually join. This research study aims to evaluate the potential of applying machine learning to enable educational institutes predict joining status of their freshmen students. Also, we attempt to understand the factors affecting joining behavior using CART algorithm. Obtaining classification accuracy up to 80 percent, it is concluded that machine learning is worth applying in this problem domain. Important factors affecting joining behavior include scholarship offered to student, fee paid so far, status of hostel facility availed and marks in qualifying examination.