Accurately predicting which students are best suited for graduate programs is beneficial to both students and colleges. In this paper, we propose a quantitative machine learning approach to predict an applicant's potential performance in the graduate program. Our work is based on a real world dataset consisting of MS in CS [Master of Science in Computer Science] students in the College of Computer and Information Science program at Northeastern University. We address two challenges associated with our task: subjectivity in the data due to change of admission committee membership from year to year and the shortage of training data. Our experimental results demonstrate an effective predictive model that could serve as a Focus of Attention (FOA) tool for an admission committee. [For the full proceedings, see ED607784.]