Authors:
S.S. Mukhil Varmann, G. Hariprasath, Irina Kadirova

Addresses:
1,2Department of Computer Science and Engineering, SRM Institute of Science and Technology, Ramapuram, Chennai, India. 3Department of Training, Silk Road International University of Tourism and Cultural Heritage, Samarkand, Uzbekistan. sm7225@srmist.edu.in1, hg8694@srmist.edu.in2, Uzpractice@gmail.com3

Abstract:

Predicting students’ performance is very important for educational institutions to get an insight into their academic progress, identify challenges faced by the students, and implement the targeted interventions for support and improvement. In this study, we propose using the H2O Gradient Boosting algorithm to predict student performance. The proposed algorithm offers several advantages, can handle large datasets, and is robust against overfitting. These features encompass academic records, socio-economic background, and behavioral attributes. We demonstrate the effectiveness of the H2O gradient boosting algorithm in accurately predicting student performance through experimentation and review. Our findings demonstrate considerable gains in predicting performance over usual techniques. The practical implications of our findings for educational institutions are substantial. So, Institutions can more effectively allocate resources to meet the unique requirements of students by utilizing the predictive potential of the H2O Gradient boosting algorithm to identify these individuals early on. It can increase retention rates, improve overall academic achievements, and create a friendly learning environment by taking a proactive approach to student support. H2O Grading is a boosting algorithm for improving predicting accuracy and facilitating data-driven educational decision-making.

Keywords: Student Performance Prediction; H2O Gradient Boosting Algorithm; Targeted Interventions; Socio-Economic Background; Resource Allocation; Data-Driven Decision Making; Large Dataset Handlin; Decision-Making Support.

Received on: 19/04/2023, Revised on: 06/07/2023, Accepted on: 05/09/2023, Published on: 22/12/2023

FMDB Transactions on Sustainable Techno Learning, 2023 Vol. 1 No. 3, Pages: 165-178

• Views : 219
• Downloads : 18

Download PDF