Authors: Kanadpriya Basu, Treena Basu, Ron Buckmire and Nishu Lal
Abstract: Every year academic institutions invest considerable effort and substantial resources to influence, predict and understand the decision-making choices of the applicants who have been offered admission. In this paper we explore and compare several supervised machine learning classification techniques to develop a mathematical model to help predict whether a student who has been admitted to the college will accept that offer. Four years of data on students admission to a small liberal arts college are used to build the classifiers and we evaluate the performance of these algorithms using the metrics of accuracy, precision, recall, F_beta score and area under the receiver operator curve (AUC). The results from this study indicate that the logistic regression classifier performs very well in determining whether the student will accept the offer extended by the college. This algorithm can prove to be invaluable by helping institutions target individuals with low chances of following through on an admission offer through emails, texts and counseling to reach target enrollment.
Details of our research results can be obtained upon request.