Automatic short answer grading

Abstract

The goal of the Automatic Short Answer Grading is to assign grades to essays and supply feedback using computers. Automated evaluation is increasingly getting used in classrooms and online exams. The goal of this research is to create and test machine learning models for doing automatic short answer grading. A publicly accessible essay data set was utilised to train and assess the efficacy of the strategies used in this study. The dataset's essays were analysed using natural language processing techniques to extract characteristics. On the specified dataset, three different machine learning techniques were applied. Among all the machine learning models, that are used in this project the sent2vec model performed the best in terms of agreement with human grading as it achieved the highest correlation value for the test dataset