Decision Tree-Based Smart Clothing in Enhancing the Psychological Care Effect of Medical Humanities Education

This study presents a smart textile system integrating an improved C4.5 decision tree for real-time identification and personalized intervention of medical students’ psychological states in medical humanities education. The system acquires multimodal physiological signals, including heart rate variability and electrodermal activity, through flexible textile sensors, and extracts temporal and nonlinear features for high-dimensional analysis. A dynamically weighted hybrid splitting criterion combining Information Gain Ratio and Gini index enhances classification accuracy and interpretability. The framework implements a closed-loop perception–decision–intervention mechanism, delivering targeted tactile feedback to regulate sympathetic–parasympathetic balance. Experimental results show recognition accuracies of 93.5% in end-of-life care and 91.8% in difficult diagnosis scenarios, with LF/HF ratio reductions of up to 46.6% and strong correlation with self-reported anxiety (R2=0.78). The approach is applicable to wearable sensing and edge-computing frameworks, leveraging real-time physiological signal acquisition and adaptive feedback. This provides a practical engineering solution for objective, timely, and quantifiable psychological care, bridging intelligent human–machine interaction with principles relevant to signal processing and embedded decision systems.