Reconhecimento Automático de Representações de Rúbricas em Agrupamentos de Soluções de Exercícios de Programação

Abstract

The evaluation of programming exercises is a complex process because, for each exercise that a teacher applies, there are common possibilities for solutions. As the teacher does not always know all the possible solutions of an exercise, it is always a challenge for him to justify all the criteria of his evaluation. In order to support the programming evaluation process, this work proposes a strategy based on clustering techniques and Principal Component Analysis (PCA) to recognize, from solutions developed by students, examples of solutions that represent, in a rubric scheme, the scores assigned by a teacher. The results of the experiments in real programming exercises solutions indicate that our method recognizes representations of rubrics demanding little teacher evaluation effort.