Toward the Automatic Correction of Short Answer Questions

Abstract

The manual correction for short answer questions is a tedious and time consuming task. This issue leads to the importance of having automatic correction systems for short answer questions. Thus this paper proposes three different approaches for correcting short answer questions. One of them is NLP-based while the others are machine learning based approaches. The three approaches are tested and compared across another two short answer correction systems that are evaluated on the same dataset. Results show that the proposed approaches outperform others that are evaluated on the same dataset.