Research On Improvement Of VaR Model Based On Brownian Motion Extreme Value Theory

VaR(value at risk)in the 1990 s as a method of measuring risk in financial markets gradually developed,its core idea is to use other standa rd statistical techniques in the field of assessment of financial risks.This method was originally applied to risk measurement,and as it was widely used and developed,it became an international standard for risk managem ent and was accepted by more and more industries and fields.With the i ncreasing risk,VaR's accurate calculation becomes the research field whic h is more concerned by scholars and financial institutions.Based on the above reasons,this paper improves the original researc h methods which only use period last time data to calculate VaR,combin es the calculation of VaR and the Brownian motion to depict the volatilit y of data for this period,and uses the optimal VaR to reduce the likelih ood of underpriced risk and get more accurate calculation method of the VaR.After obtaining the theoretical model,this paper uses Monte Carlo si mulation method and R software to test the rationality of the minimum v alue distribution obtained using the Brownian motion extreme value princi ple.The test results show that the minimum value distribution is reasonab le,indicating that the improved model has theoretical significance.The e mpirical analysis section uses the failure test in the Kupiec test to verify the optimization effect and robustness of the improved model.This sectio n compares the empirical results of the modified model with that of the t raditional model.The results show that the model has a good improvement and can accurately estimate the stock market's trend.