DEVELOPMENT OF COMPUTER BASED ANSWER SCRIPT EVALUATION SYSTEM
No Thumbnail Available
Date
2023-03-15
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Computer-based evaluation of answer scripts is becoming increasingly important in
the field of education. In this thesis, we propose a computer-based answer script
evaluation system that evaluates both text and figures in answer scripts. For text
evaluation, we consider two cases and use different methods for each case, including
writing score, keyword matching score, Cosine similarity score, NLP-based cosine
similarity score, and plagiarism matching score. Our system achieves high accuracy
scores of 83% and 85% for each case, respectively. For figure evaluation, we use
cosine similarity and Euclidean distance techniques, achieving accuracy scores of 85%
and 72%, respectively. Our system provides a robust and reliable way to evaluate
answer scripts in an automated manner. By using a combination of different evaluation
techniques, we are able to effectively evaluate both text and figures, improving the
accuracy of the overall evaluation. This system has the potential to greatly reduce the
workload of educators, while also providing students with more immediate feedback
on their work. Overall, our system represents a significant step forward in the field of
automated answer script evaluation.