Research On The Theory And Realistic Meaning Of Hidden Markov Model In The Algorithm Trading

With the development of the Chinese Stock market,researches have spring up continuously in both theory and practice.Algorithm Trading is a trading method with knowledge of both Finance and Computer Engineering.From the perspective of the Algorithm trading,Stock Market lies between inefficiency and the weak-form efficiency:Not all the historical information can be included in the stock price,say,the psychological condition of the traders is invisible.With the preprocessing of the market data,some indicators,can be transformed to the trading signal.Thus,it is more effective to use Computer to analyze and manage those trading signal.Hidden Markov Model(HMM)is a framework about Stochastic Process.With some reasonable restrictions,it made the complete stochastic process to be a quantitative model,so that people could explain it with the help of mathematic.From this point of view,the simulation with the HMM is just a market between inefficiency and the weak-form efficiency.HMM is a dual stochastic process:The Hidden State transforms with the change of time series.Such transform as well as the Hidden States cannot be observed.They can be speculated only by the Observation,where Observation come from the Hidden State.HMM reflect the mechanism of the Stock Market to a certain extent:The rise and fall of a stock is invisible——what is observable is the stock price,the trade volume.In this paper,a Hidden Markov Model with the Time Series would be studied in order to forecast Stock Market in Algorithm Trading.Chapter II explained the theoretical fundament of the Hidden Markov Model with a time series analysis,as well as its origin,the Bayes Theory.Chapter III gives a concrete operating mechanism and also tests the common technical factors.Chapter IV presents the summary and the advancement in the future.