The Markov Model And Application Based On Theory Of Evidence

The Markov chain is an important part of the random process theory,and has a variety of applications[7-13].The classical Markov chain model is mechanical and single in the determinable process of the state transition matrix.The theory of evidence is more effective than the theory of probability in expressing and solving the problem of uncertain information.This theory has been widely recognized by many scholars and used in various fields.Xinyang Deng[6],the first person,attempted to introduce evidence theory to improve the state transition matrix of Markov model.This paper based on this idea proposes an improving method of transition matrix in the states.get the reliability function and the corresponding value by introducing the D-S theory of the original data,and gives a continuous state function value method to determine the general reliability;through the method of deviation square and weighted backward based on the reliability function value of the weight distribution form transition matrix,the transition matrix considering the cumulative effect between states.Finally,through the example analysis,the comparison between the different models under the state of the back substitution error,obtains the good forecast result.Hidden Markov model is based on Markov model is the premise and an extended statistical analysis model is a certain number of states Markov chain process and double stochastic process stochastic observation function set.In thi s paper,we give a detailed solution to the three classical problems of HMM and give a concrete example.Simultaneous,combined with the theory of evidence and the method of probability transformation,a new method for improving the transition matrix of hid den Markov model is proposed.Then using the 2005-2016 stock market SSE Composite Index as an example,the use of the credibility of the weighted Markov model and analytic hierarchy process,proved that the stock market trend is a cyclical process.Final ly,we use the hidden Markov model to predict the stock market trend for 9 consecutive days.