Constructing A Programming Trading Strategy Based On Online Trading Theory And Hidden Markov Model

In recent years,securities trading has been developing towards diversification,complexity and huge volume of transactions.Traditional manual ordering has long been unable to meet the needs of the market,and investors are turning to programming trading way.Since the advent in the 70 s last century,programming trading has become one of the main trading methods in the international securities market after more than forty years' development.This kind of trading method can not only improves the income for investors,but also accelerates the liquidity of the financial market.At present,the application of programming trading has been very mature in foreign market while is relatively lagging behind in China.Since the programming trading is still unfamiliar to most of the investors,it is of great practical significance to help our investors to design programming trading strategies.In order to enrich and improve the trading strategy in China's stock market,based on the A share market in China,we design a new type of trading strategy by combining the two online trading theories-universal portfolio-and K search algorithm,and hidden Markov model recognition model.Firstly,trading strategy EGK is designed according to the universal portfolio strategy EG(?)(Exponentiated Gradient)and the K-Search algorithm.Then,noticing the distribution of stock returns rate in China A stock market-China A stock market has not yet reached the weak efficiency and thus the stock price has correlation-the strategy EG(?)was improved to strategy IEG(?)(Improved EG(?)),and thus EGK(EG(?)and K-Search)is improved to IEGK(Improved EG(?)and K-Search).which is tested on the quantitative trading research platform of Uqer,showing that the strategy IEGK has stable profitability in Chinese A stock market.The relationships between the three parameters and the rate of return about strategy IEGK are also discussed in detail,respectively,deepening the understanding of the applicable scenarios of strategy IEGK.Finally,when discussing the parameters of strategy IEGK,it is found that the anticipation of market trend will significantly affect the profit of strategy IEGK,who enlightens us to do market timing research on market state prediction.By introducing pattern recognition to the hidden Markov model,the trend of the market anticipation can be predicted,and thus the optimal parameters of trading strategy IEGK can be determined,causing our design of HMM+IEGK(Hidden Markov Model + Improved EG(?)andK-Search).Since the empirical results show that the strategy HMM+IEGK can greatly improve the performance of strategy IEGK,we claim that HMM+IEGK is a more effective program trading strategy.Programming trading strategies designed in this paper have both theoretical and practical value.They are importance references for the A stock market investors in improving the weak investors' investing capabilities,and thus can promote the development of fairness in China's stock market.