Research On High Frequency Trading Strategy Based On Maximum Entropy Spectrum Analysis

In recent years,with the deepening of China's financial reform,financial investment demand is rising,network equipment and hardware level increasing,investment strategy,especially the high frequency trading strategy is of particular concern by the market participants and managers.High frequency trading is a form of dynamic trading,known for its speed,it the use of computer technology and system,the execution of the transaction at high speed,and the days of short positions.Normally,T+1 market is not in high frequency trading,so Chinese focused on high-frequency trading futures(gold in the stock index futures or the futures exchange of commodity futures,margin trading)ETF and other fields.High frequency trading has become a mainstream means of trading patterns in the United States and Europe,its profit model is completely different from the general behavior of traders,the trading profit pattern regardless of trading psychology and fundamentals,often by means of computer program development,designed trading strategies,which will then be compiled into a computer program.The application interface for exchange of orders and return because of judgment,the execution of the strategy is based on the computer program algorithm to determine the timing of trading orders,price and quantity,so compared to manual orders,high frequency trading program in the management of the market impact cost.It has a series of advantages,opportunity cost and risk.It can be divided into common high-frequency trading strategies(1)arbitrage strategies,including arbitrage,cross species arbitrage;(2)the short-term speculative strategy,according to the price and trading volume,or technical indicators of information statistics results of the calculation parameters,decision open or closed time,micro application transactions model tracking a trend in a short time interval.The high-frequency trading strategies profit space is established in the rapid capture trading signals,as well as the basic transaction fees and low effective.One for the trend of short-term speculation in the market tracking strategy,huge fluctuation,the trend is good,trading signals effectively and winning high,but when the market turned into concussion,trading signal is the error signal resulting in consecutive losses,the maximum retrace also generated in this stage.Therefore,if we can effectively identify trends,grasp the trend,and shield the shock period,it has a greater chance to obtain speculative gains from it.In this study,according to the theory of maximum entropy spectrum estimation,time domain data using innovative method for transform frequency domain data to design transaction model with shocks and trend identification ability,and gives the parameter tuning method to estimate the model.At the same time,on the basis of relevant theory,in order to the effectiveness and operability validate true,the historical data using multiple market many varieties of back test analysis,according to the characteristics of back testing after the transaction results and actual profit distribution,analyzes the method for various subject feasibility,put forward countermeasures and suggestions on further reform in unilateral speculative trading strategy in China market.