Research On The Establishment Of Risk Early Warning System Of Chinese Stock Market Based On BP Neural Network Optimized By Cuckoo Search Algorithm Research

With the deepening of global economic integration,the trend of financial market globalization is becoming more and more obvious.Looking back at the major financial crises in history,they not only caused heavy losses in the countries where the crises occurred,but also had a serious negative impact on the global financial market and the world economy.Therefore,risk prevention and control work is the top priority in the financial industry in recent years.China's stock market started late,but developed quickly,which promoted the rapid growth of China's overall economic level.But the stock market is a typical high risk,high yield market,for our country,is a new,rapid development and immature market.Therefore,it is very important to select effective risk warning indicators and warning models to build an effective stock market risk warning system to monitor and analyze stock market risks,so as to avoid the outbreak of market crisis,safeguard the rights and interests of investors and the healthy development of financial market.Nowadays,the extensive research and application of artificial neural network has gradually developed into an international frontier research field,providing a new research method for risk early warning.In this paper,the research on stock market risk early warning index,early warning model and neural network optimization algorithm by domestic and foreign scholars is reviewed.On the basis of analyzing the risk incentive of China's stock market,a stock market risk early warning system based on double-hidden layer BP neural network optimized by cuckoo bird search algorithm is proposed.The specific research contents and results are as follows:(1)Based on the research of existing literature,this paper selects 18 representative indicators from five aspects of stock market risk incentives:domestic economic situation,international economic situation,internal risk of stock market,safety of financial institutions and investor behavior,and constructs an early-warning index system of China's stock market risk.The empirical results show that the index system can play an early warning role on the risk of China's stock market.(2)In terms of the selection of early warning model,this paper selects the BP neural network which is widely used in the early warning model of financial system,and tries to make a new attempt for the early warning system of stock market risk in China,and optimizes it with cuckoo search algorithm in view of its vulnerability to local minimum.At the same time,in order to improve the prediction accuracy,this paper adopts the double-hidden layer BP neural network model optimized by cuckoo search algorithm to study the risk early warning of China's stock market.Through the analysis of the results of model training and test,and the validity test of the established model based on China's two stock market crises in 2008 and 2015,it is found that the early warning results of the model conform to the actual risk situation of the stock market.Therefore,the early-warning model of stock market risk established in this paper has good early-warning effect.