Optimizing Predictive Model for Student Programming Ability using Regularized Multiple Regression Analysis

Predicting the programming ability of a student involves assessing various aspects of their coding skills and problem-solving capabilities. By carefully designing and utilizing rubrics, instructors can gauge a student's programming score based on specific criteria. In this research, a regularized predictive model is proposed to predict the programming ability of a student using rubrics generated score. Rubrics are an essential tool for assessing student programming assignments, providing a structured and transparent method for both teachers and students. Rubrics outline the criteria for evaluation, clarifying expectations for code functionality, style, and documentation. This helps students focus on the key aspects of the assignment. In this research, initially, a dataset is compiled having different features to calculate the score of the students. In the second step the regression analysis is performed first to predict the student. By implementing a clustering based automated feedback system using the k-mean algorithm, instructors can provide more timely and detailed feedback, which can significantly improve student learning outcomes in programming courses.