A study on China's "silver economy" based on SHAP interpretability modeling

This paper aims to explore the development models of China's provincial silver economy, identify their core influencing factors and interaction effects, so as to provide a basis for formulating targeted policies. To achieve this goal, it selects the panel data related to the silver economy from 2014 to 2023 in 31 provinces, municipalities and autonomous regions across the country, constructs a measurement index system including population structure, uses the EWM-CRITIC coupled weight system to measure the development level of the provincial silver economy, divides three types of models through spectral clustering and analyzes their characteristics and causes, and then uses random forests and SHAP method in explainable machine learning to identify the influencing factors and interaction effects. The results show that the provincial silver economy can be summarized into three types of models: high population structure-high consumption capacity type, low consumption capacity-medium pension security type, and low population structure-low elderly service resources type. Pension security and elderly service resources are the core influencing factors of the silver economic development model, and the population structure and consumption capacity constitute the demand base, and policy subsidies play a regulating role. The research innovation combines EWM and CRITIC methods to determine weights, and uses SHAP algorithm to analyze nonlinear effects and interaction effects, which can provide a basis for local governments to formulate policies, enterprises to tap potential, and the elderly to enhance consumer confidence.