Key Factor Analysis of Digital Health Ability in the Elderly Using SVM and Shapley Algorithm

Abstract

This research centers on the analysis of critical factors influencing the digital health ability of older adults in a city during the post-pandemic period, addressing the current situation of low digital health proficiency and a pronounced digital divide in a specific country by developing a comprehensive evaluation framework. This study employed a SVM model with K-fold cross-validation for optimization and utilized the Shapley algorithm to quantitatively assess the contribution of critical factors to digital health ability. The findings indicate that technology anxiety, community support, and mental health are the primary factors affecting the digital health ability of older adults, with technology anxiety exerting a pronounced negative influence. The optimized SVM model showed substantial improvements over traditional stepwise regression models regarding predictive accuracy and robustness, highlighting its strengths in capturing nonlinear relationships. This research offers a scientific foundation for policymaking, stressing the importance of alleviating technology anxiety, strengthening community support, and enhancing mental health to narrow the digital divide among older adults and elevate their digital health literacy comprehensively. Furthermore, the approach and findings of this study are applicable to other areas of health management and digital literacy research, offering robust support for the advancement of healthy aging and the realization of social equity.