Learning Behavioral Characteristics and Teaching Intervention Strategies of Midwifery Students in Higher Vocational Colleges Using K-means Clustering Algorithm

To address the difficulty of integrating multi-source heterogeneous learning behavior data and the broadness of conventional teaching intervention strategies, this paper proposes a data-driven analysis framework based on K-means clustering for higher vocational midwifery education. The study collects multidimensional behavioral data from online learning logs, smart classroom interaction records, and practical training assessments, and constructs a 12-dimensional feature system covering cognitive input, behavioral activity, practical competence, and comprehensive performance. Zscore standardization is used to eliminate dimensional differences, and the optimal cluster number K=3 is determined by combining the elbow rule and silhouette coefficient. Students are classified into actively balanced, theoretically weak, and practically lagging groups. Experimental results show that targeted interventions based on clustering profiles significantly improve core competency indicators in theoretical and practical dimensions. Hotelling’s T2 tests and external evaluation indicators confirm the statistical significance and educational interpretability of the clustering structure. The framework can also support engineering training environments using smart classrooms, wireless sensing, and electromagnetic-compatible data acquisition systems for precise learning diagnosis and intervention.