Analysis of Regional Financial Risk Factors under Random Forest

Objectively understanding the current state of China's financial risks and their external drivers is a crucial guarantee for maintaining financial security and stability, and promoting high-quality financial development. This paper breaks through the limitations of traditional indicators in monitoring shadow banking, cross-border capital, and the marginal effects of debt risk. It employs the entropy weight method to construct a comprehensive regional risk assessment framework covering four sectors, and empirically analyzes the main external influencing factors of regional financial risks based on the random forest model. The research indicates that: (1) China's regional financial risks exhibit significant spatial differentiation, with major influencing factors forming a four-tier risk driver system. (2) The risk sources in different strategic regions stem from the reshaping of regional economic structures due to national strategic positioning. This is manifested in three regional dynamics: the Beijing-Tianjin-Hebei region demonstrates dual constraints through state-owned enterprises' dominance in major project investments; the Yangtze River Delta, as an export-oriented economic hub, exhibits heightened susceptibility to international market fluctuations; while the Yangtze River Economic Belt illustrates the critical equilibrium between large-scale infrastructure investment and local fiscal capacity in cross-regional coordination. (3) Regional risk governance needs to emphasize the asymmetry of dynamic evolution, providing a scientific basis for differentiated policy measures.