Font Size: a A A

Parameter Estimation Based On Optimal High Frequency Trading Of Stock Index Futures

Posted on:2020-08-15Degree:MasterType:Thesis
Country:ChinaCandidate:S WangFull Text:PDF
GTID:2439330575480394Subject:Probability theory and mathematical statistics
Abstract/Summary:PDF Full Text Request
With the development of financial marketing,the ratio of high-frequency trading increases a lot.Studying the application of high-frequency trading in China's market is of great significance to the study of local financial market.Introducing the high-frequency trading strategy in the foreign stock market into local current stock index futures market can not only introduce the market maker system in local current bid trading mechanism but also provide very important opinions and suggestions for local market operation.Fabien et al.used the Markov chain to simulate bid-ask spreads in research and the intensity of execution differed when speculators chose different offers.The speculator can submit a limit buy order(or sell order)at the current optimal bid at any time but must wait for the market order that matches its limit order to enter the marketplace for execution.The speculator can also choose to buy(sell)at a slightly higher(slightly lower)price than the current optimal bid,thereby controlling the adjustment of his bid and offer and executing the trade faster.Through the study of the intensity of execution,weigh the market price directive and the limit order,so as to maximize the expected benefits in the short term.Combining optimal control problems with dynamic programming,the market-making problem is reduced to a dynamic programming problem that considers only two variables,inventory and spread.Based on the previous studies,this paper studies the optimal high-frequency trading strategy when speculators in China's stock index futures trading use the difference between high-frequency trading of stock index futures to earn profits.An empirical study was carried out on the execution point process and the parameter estimation of the difference based on the high-frequency trading data of CSI 300 stock index future.In the parameter estimation of the difference,Fabien's estimation method is used to estimate the difference parameter of the high-frequency trading data of CSI 300 stock index future.Because of its tendency to exhibit L-shaped and U-shaped curve,we consider the hyperbolic function to fit the parameter estimation of the difference,and the time interval is refined to observe the fitting effect,but the effect is not good.Furthermore,we consider the parabolic function to estimate the difference.The effect is better than the hyperbolic function fitting,but there is still room for improvement.Therefore,we consider the estimation of the variation intensity function of the spread jump process,the accuracy and fitting effect of the estimation are improved.The maximum likelihood estimation combined with the parabolic function fitting method is used to improve the parameter estimation method.The fitting effect is better than before.The maximum likelihood estimation with better parabolic function fitting is proposed,in order to obtain the optimal high-frequency trading strategy suitable for local stock index futures market,which provides a new idea on application for the optimal high-frequency trading strategy in local stock index futures market.
Keywords/Search Tags:Limit order, market order, point process, stochastic control, estimation
PDF Full Text Request
Related items