A Machine Learning Approach for Quantifying Academic Misconduct

Evidence from the literature continues to reveal the problem of academic misconduct, particularly cheating, among university students. To deal with this problem effec tively, a clear understanding of its magnitude is necessary for planning and resource allocation. This paper proposes a machine learning algorithm to quantify the mag nitude of academic misconduct among undergraduate students. In this study, cluster analysis was employed with outlier detection and removal. The algorithm was trained on a dataset comprising 678 short texts. Results indicated that over 80% of students engage in the practice of academic misconduct. This shows that academic misconduct among undergraduate students poses a serious risk to the quality of graduates. This paper proposes a machine learning algorithm to quantify academic misconduct. The proposed algorithm is based on a modified k-means clustering algorithm that auto matically detects and removes outliers. Universities can adopt the proposed method to combat the growing problem of academic misconduct among undergraduate stu dents. The proposed approach for quantifying the magnitude of academic misconduct is more reliable and cost-effective than traditional (survey-based) methods.