NLP-based Automatic Answer Evaluation
Shubham Kumar Sinha, Sachin Yadav, Bindu Verma
Abstract
Using answer scripts to evaluate students is an important part of evaluating their performance. In most cases, an answer script evaluation is done dynamically, which might head to prejudice. It depends on a number of factors, including the mood of the teacher and the interaction between the student and the teacher. Further, checking is an important part of the process an extremely time-consuming and arduous process This study demonstrates a natural language processing-based technique for evaluation of answer scripts by an algorithm. To score the answer script, text is extracted from the answer, the extracted text is compared with the stored correct answers to calculate similarity, then a weight value is assigned to each measure. Keyword-based summarizing algorithms are utilized to generate a summary from the retrieved material. For producing the final mark, four similarity measures (Cosine) are employed as parameters. These investigations have shown that automatic assessment of response scripts is quite valuable, and that the given marks are often as same as the hand marks scores.