GRAD-AI: An automated Grading tool for code assessment and feedback in programming course

Abstract

The current educational system relies heavily on manual grading, posing challenges such as delayed feedback and grading inaccuracies. Automated grading tools

(AGTs) ofer solutions but come with limitations. To address this, "GRAD-AI" is introduced, an advanced AGT that combines automation with teacher involvement

for precise grading, timely feedback, and personalized support, enhancing the education process. This paper aims to develop "GRAD-AI," an automated grading tool for

computer programming assignments. The objective is to overcome the limitations

of manual grading by harnessing AI’s capabilities to deliver accurate and timely

assessments, thus creating a more interactive and supportive learning environment.

The results show that GRAD-AI provides unbiased grading and timely and accurate

feedback delivery for programming assignments by using the Halstead Complexity

Measure, Term Frequency – Inverse Document Frequency Measure, Abstract Syntax Tree Process, and K-means Clustering. GRAD-AI marks a substantial stride in improving grading and feedback delivery within the education sector. Its real-time

feedback provision and gap identiication contribute to enhanced learning outcomes. As AI’s role expands, integrating automated grading tools like GRAD-AI becomes crucial for fostering personalized learning and adaptability. The paper underscores

AI’s potential to revolutionize assessment and grading processes, supporting global

students’ growth and development. Keywords Automated grading tool · Feedback · Programming assignments · Assessment · Automated · Manual · Python programming · Artiicial intelligence · Characteristic extraction · K-means clustering · Plagiarism detection · User Interface