Development of Automatic Answer Evaluation System using Synonym Method

One of the most crucial phases of the education system is the assessment of student responses. Automatic assessment of the students' responses has become one of the most crucial tasks resulting from the evolution of natural language processing. In this paper, we bring out a method for the use of WordNet to evaluate answers based on synonyms. WordNet is a lexical database of semantically related English words. In the proposed method, WordNet is used to find similar terms in the student's response by matching synonyms of the student's response against synonyms of the correct answer. After that, the highest resembling synonyms are picked, and a similarity score is calculated. Results indicate that this approach can yield reliable and uniform evaluation results, apart from high ability in the evaluation of answers based on synonyms. The technique yielded reliable and consistent evaluation results, with the overall similarity ratings from each question showing the following: 74.40%, 89.60%, 92.20%, 94.60%, and 83.60%.