Combining Machine Learning and Multicriteria Decision Support in the Evaluation of Graduate Courses in Brazil: A Case Study in the Area of Computer Science from the CROWM method

This study presents an innovative methodology to evaluate Brazilian postgraduate courses in the 2013-2016 quadrennium using the Communalities on Ranking and Objective Weights Method (CROWM). Integrating Machine Learning and Multicriteria Decision Making, the objective is to elaborate, through a case study, a methodology capable of evaluating 17 graduate courses in Computer Science based on six performance criteria taken directly from official information from the Coordination for the Improvement of Higher Education Personnel (CAPES): Planning, Infrastructure, Quality of Teachers, Quality of dissertations, Efficiency, and Quantity of Productions. Reaching a Pearson correlation of 0.9 between the ranking generated and the CAPES Scales of each course evaluated demonstrates the method's robustness. In addition, by helping to identify strengths and opportunities for improvement, the results of this study can contribute to assist in the improvement of programs in pursuit of academic and scientific excellence.