Research on the Analysis of Potential Abnormal Student Behavior in Psychological Assessment Based on K-means Clustering

Abstract

This study is based on ideological and political education, and proposes a potential psychological assessment abnormal student behavior analysis model based on K-means clustering to address the psychological health management needs of college students. By integrating psychological assessment scale data (such as SDS and UPI scales), student abnormal behavior, and gender indicators, a multidimensional characteristic psychological health education evaluation system is constructed; And the K-means clustering algorithm is used to identify optimal cluster centroids, combined with different contour coefficients and elbow rules to determine the optimal number of clusters, achieving automated recognition of abnormal behavior groups. Experiments have shown that when the K value is 4, the SC value is greater than 0.5, indicating that the model has a certain degree of accuracy and scientificity in identifying high-risk groups such as depression tendencies and suicide situations; The clustering results can effectively classify abnormal risk situations into normal, mild, moderate, and severe, providing data support for precise intervention of abnormal students and a new paradigm for digital governance of mental health in universities. Meanwhile, this framework enables proactive, data-driven mental health interventions, promoting early prevention and personalized psychological support in universities.