Information supervisory model for financial risk prevention and control based on twin-SVM

Aiming at the problems of poor effect and low accuracy of traditional financial market risk prevention and control methods, a financial risk prevention and control information monitoring model based on double support vector machine is proposed. The improved support vector machine and asymmetric COvAR data were combined with covariance operation to reduce the actual tail risk overflow. Through financial aggregation and covariance tail data of current characteristic financial system, the spillover effect of financial risk is obtained. According to the extreme value statistical analysis theorem, it is determined that the current financial risk gradually obeys the extreme value. In order to verify the effectiveness of the method, the financial data from 2012 to 2018 provided by a bank were used as experimental samples to conduct simulation experiments. Experimental data show that the proposed method has lower boundary cost and higher market arbitrage, and has a strong market applicability.