An LSTM And BiLSTM Models for Automated Short Answer Grading: An Investigative Performance Assessment

Please use this identifier to cite or link to this item: http://repository.futminna.edu.ng:8080/jspui/handle/123456789/27538

Title:	An LSTM And BiLSTM Models for Automated Short Answer Grading: An Investigative Performance Assessment
Authors:	Nusa, Aisha Muhammad K. Bashir, Sulaimon Adebayo Adepoju, Solomon
Keywords:	Automated Short Answer Grading; Bidirectional LSTM; Deep learning; LSTM
Issue Date:	Mar-2023
Publisher:	4th International Engineering Conference Federal University of Technology, Minna Minna, Nigeria.
Citation:	31. Nusa, A.M.K, Bashir, S.A., & Adepoju, S. (2023). AN LSTM AND BiLSTM MODELS FOR AUTOMATED SHORT ANSWER GRADING: AN INVESTIGATIVE PERFORMANCE ASSESSMENT. Proceedings of the 4th International Engineering Conference Federal University of Technology, Minna Minna, Nigeria
Abstract:	Automated Short Answer Grading (ASAG) systems contributes immensely in providing prompt feedback to students which eases the workload of instructors. In this paper, the performance of two deep learning models (LSTM and BiLSTM) were investigated to ascertain their effectiveness in grading short answers. The popular ASAG dataset by Mohler was utilized for the experiment. The dataset contains training samples from Computer Science department with grades between 0-5. The results show that LSTM model performs better interms of training time with lower RMSE and MAPE when compared with BiLSTM.
URI:	http://repository.futminna.edu.ng:8080/jspui/handle/123456789/27538
Appears in Collections:	Computer Science

Files in This Item:

File	Description	Size	Format
Nusa conference.pdf		2.22 MB	Adobe PDF	View/Open