| With the rapid development of science and technology in China,the research and application of intelligent system are more and more extensive.As the main body of quantitative investment in financial market,intelligent trading system has developed for more than 30 years abroad.At present,the field of domestic financial investment still have broad development space in China.The intelligent trading system carries on the deep research to the historical big data by the powerful computer function,optimizes the strategy and the parameters,establishes the best strategy model suitable for various varieties,and may be effectively verified and perfected in the actual trading situations.Moreover,the intelligent trading system has higher profitability and execution efficiency than artificial subjective operation,so the research of financial intelligent trading system is particularly important.As a typical complex system,the fundamental significance of the research on financial system lies in the realization of statistical prediction.And the conditional statistical properties which is the most important statistical property of complex system,is also the core theoretical foundation of the realization of statistical prediction in financial system,and construct the intelligent trading system.The main contents of this thesis are as follows:1.Taking the commodity futures rebar's price system as the research object,the research selects the data period from March 27 th 2009 when the rebar futures is officially listed on the Shanghai Futures Exchange to October 30 th 2017,with a total of 585395 high-frequency trading minute-by-minute data.By means of statistical physics,the dynamic model of the rebar's price system is established,then the anomalous statistical property and the autocorrelation property of the system are studied,and the financial price system's distribution can be well fitted by q-Gaussian shape is proved.2.This research process the minute-by-minute data of rebar,establishes the ideaof the moving average prices,and proves that the return rate of the moving average prices in the system has well positive correlation property.According to the distribution form of q-Gaussian,the conditional probability distribution function of the moving average prices system can be obtained by derivation.On the basis of this,the concept of threshold is introduced into the conditional probability distribution function to construct the statistical prediction model of the conditional threshold of the moving average prices.3.According to the theoretical model of the conditional threshold of the moving average prices,the corresponding intelligent trading system is designed.At the same time,the program of historical data back testing is designed to verify the simulated trading results of the intelligent trading system,and provides a scientific basis for the combination of transaction indicators.We present and analyze the results of simulation transaction from October 31 th 2017 to March 27 th 2018.The final transaction results show that the intelligent trading system designed in this research have the advantages of stable profitability and anti-risk capability.It may be proved that the intelligent futures trading system based on the conditional statistical properties of the moving average price is reasonable,effective and practical.The research work in this paper is not only of guiding significance to the theoretical study of the conditional statistical properties of complex system,but also of practical significance to the practical application of the financial intelligent quantitative trading system. |
PDF Full Text Request