Studies On The High Frequency Trading Risk Forecast Of Stock Index Futures Based On The VPIN Model

With the development of technology,high frequency finance emerged.At the same time,the high frequency trading gradually increased,and the problems of low frequency model failure,high frequency data noise,and transaction timeliness were also faced.The presentation,analysis and prediction of high frequency trading risk are the key to solve the above problems.This paper is based on the comparison of high-frequency trading risk agents and related prediction models.This paper uses one minute data of stock index futures of the IF300,IH50,IC500.High frequency liquidity,absolute return volatility,relative return volatility to the earnings volatility as the proxy variable,multiple prediction model is given by using the combination of multiple independent variables.The analysis process introduces VPIN as external information to reduce the influence of microstructure noise.For the accurate estimation of the VPIN model,the most applicable high frequency transaction driving direction identification algorithm is also analyzed.Based on the analysis results,this paper obtained the high-frequency trading risk prediction model,and used the indicators such as VPIN to predict the high frequency trading risk.First of all,in order to accurately measure the VPIN model,in view of the high-frequency drive direction recognition algorithm selection problem,this paper presents a theoretical analysis based on bayesian learning model,marking rules classification algorithm and applied conditions of total volume classification algorithm,further use of actual high-frequency trading stock index futures market data,test current is more applicable to Chinese stock index futures market trading drive direction recognition algorithm is total volume classification algorithm.This provides support for the next VPIN model measure.Secondly,in order to solve the problem of the daily volume basket number of the VPIN model parameter,this paper takes the intraday VPIN value as the criterion and selects the parameter N=20 with the highest correlation.In has been clear about the number of average daily turnover basket problem,on the basis of using the VPIN model of IF300,IH50,IC500 stock index futures market to measure the probability of informed trading,and the results are the descriptive statistical analysis,sample period price trend analysis and comparison of prices before and after the stock index futures incident analysis;The research shows that the VPIN method has good applicability in China stock index futures market and provides support for the next step of high frequency trading risk forecasting.Finally,for the risk of high-frequency trading,this paper comprehensively takes into account the level of liquidity and volatility,and selects three high-frequency trading risk indicators and measures.In the measure of high-frequency trading risk problem,on the basis of establishing high frequency scenario liquidity level,the level of volatility and the issue of market between the probability of informed trading model and parameter estimation and inspection,studies have shown that VPIN method can in China's stock index futures market specific scene play a good role in warning.In view of the high frequency trading risk prediction question,this paper researched based on high-frequency trading.The combination of theoretical innovation and practical application,from two aspects: mathematical modeling and empirical research,promote the high-frequency trading risk research in depth and breadth.