A Study of Factors Affecting the Students' Computer Programming Achievement Using Machine Learning Technique

This is the authors' abstract. We don't add key points for this paper.

Acquiring computer programming skills frequently poses difficulties for students because it necessitates analytical reasoning, logical thinking, and problem-solving abilities. The challenges of mastering programming syntax and debugging errors contribute to elevated dropout rates and diminished engagement. This research utilizes machine learning methodologies to forecast students' academic performance in computer programming and discern critical factors influencing their success. A Support Vector Machine (SVM) model was employed in two iterations of analysis involving 750 records. The initial round encompassed all 29 attributes, whereas the subsequent round concentrated on five principal attributes determined via multiple regression analysis. The research identified five principal attributes—high school GPA, self-directed learning attitude, experience with programming flowcharts, learning style preference, and career motivation—as significant predictors of learning outcomes in computer programming, attaining an accuracy of 72.40%. This allows educators to customize approaches, including the simplification of intricate subjects, the integration of experiential activities, and the provision of additional resources such as exercises and tutorials, to enhance student.