Automated Grading and Feedback Systems for Programming in Higher Education Using Machine Learning.

Abstract

Evaluating programming tasks in higher education is difficult, sometimes characterised by inconsistency and insufficient feedback, hence constraining student development. This study introduces an automated grading and feedback system powered by machine learning to tackle these difficulties. The system utilises supervised learning to forecast grades with 98.5% precision, including test case analysis, structural validation, and natural language processing for feedback production. The suggested methodology offers enhanced precision (97.8%) and recall (98.3%) relative to existing methods, guaranteeing grading accuracy and constructive feedback. The results indicate its efficacy in managing extensive submissions with minimal interruption, providing scalability and stability. The research considerably improves the grading process in higher education by mitigating the drawbacks of previous techniques, including prejudice and inefficiency, hence promoting superior learning results. Future initiatives involve enhancing support for multi-language programming and improving feedback mechanisms to provide adaptation across various educational environments.