Research on the Optimal Action Time Window of Financial Support Policy Based on Deep Learning

As China's financial market opens up and innovation accelerates, regional financial risks have gradually emerged and become one of the key concerns of current financial regulation. This project takes Liaoning Province as an example, takes the spatio-temporal correlation between financial regulation and economic development as the research object, and selects a total of 16 mixed-frequency sample indicator pools representing the state of economic and financial development from five dimensions. Recurrent neural network (RNN) in deep learning is used as the basis for model construction to fit the trend of GDP under different circumstances. The effectiveness of regional financial regulation policies is explained from a quantitative perspective. The optimal time window for the introduction of financial support policies is found through model generalization. This research will play a crucial role in systematically preventing regional financial risks, giving full play to the incentive effect of policies, and better leading economic development.