Automatic Recommendation of Students' Answers for Error Mediation

Resumo

The problem of recommending answer records for error mediation in educational environments is introduced. Teachers face several difficulties in analysing students incorrect answers due to the high workload it requires, even if these answers are digitally stored. In order to provide and facilitate the error mediation, this study describes an algorithm for answer recommendation. This algorithm generates automatic recommendations of relevant answers and questions to groups of similar students. Preliminary tests in a real application have indicated that the algorithm is capable of defining groups of students with actual similarities on their errors and of generating relevant recommendations. The main conclusions of the study are described and future works are pointed out.