<![CDATA[This study profiles middle school students’ participation in environmental conservation using K-means clustering as a data-driven approach to identify latent behavioral patterns. A total of 457 students participated, and their responses were analyzed across four dimensions: participation frequency, environmental knowledge, attitudes and behaviors, and social involvement. The optimal number of clusters was determined through elbow and silhouette analyses, with the three-cluster solution offering the best balance between statistical coherence and interpretability. The clusters revealed distinct profiles: a high-engagement group with strong environmental literacy, a moderate group exhibiting a knowledge–action gap, and a low-engagement group with limited awareness and participation. Kruskal–Wallis tests confirmed significant differences across all dimensions (p < 0.001). Principal component analysis (PCA) visualization further supported the interpretive clarity of the cluster structure. These findings demonstrate the heterogeneity of students’ environmental engagement and highlight the importance of differentiated environmental education strategies. The study underscores the value of integrating machine learning analytics with pedagogical design, offering actionable insights for schools to strengthen conservation education through targeted experiential activities, leadership opportunities, and foundational environmental support for the least engaged learners]]>
Copyrights © 2026
