Automated Explanatory Answer Evaluation Using Machine Learning Approach

Abstract

Assessment plays a critical role in a learning process to evaluate the students’ knowledge regarding the concepts related to learning objectives. The current pandemic situation forced the educational institutions to adapt online education followed by online evaluation. Online evaluation will be comfortable with multiple choice questions compared with evaluating explanatory answers. The online evaluation of explanatory answer is a difficult and time-consuming process since experiencing discomfort in their eyes, unequal assessment even for the same answers due to depression and less interest in evaluation.

Hence, this paper is focused on the development of evaluation of explanatory answer using machine learning approach. Initially, Natural Language Processing(NLP) is used to extract the keywords from the answer scripts as well as from the answer key using the tasks like Tokenizing words and Sentences, Removal of Stop words and Lemmatization Process. Finally, to determine the degree of similarity between the students answers and answer_key, the proposed system uses Cosine Similarity as the similarity measures. The mark is assigned to the answer scripts based on the criteria and the similarity measures value. The proposed system is trained and 100 different students’ answer scripts are evaluated using machine learning approach.