Moving Average Markov Model Of The Highway Along The Rainfall Forecast

The moving average Markov model is based on Markov model, which deal with by the moving average before use Markov model to forecast. It offers a new path to forecast problem and enriches the content of forecast. The basic way to forecast is:Firstly, the dates which were dealt with by the moving average, reduce the effect by the accidental factor.Secondly, the phenomenon is looked upon as a system, and partitioned to several states.Thirdly, transition probability is worked out.Fourthly, how the system shift is confirmed.Finally, based on the transition probability and random matrix, the probability of several shift is speculated. The forecast states are changed to a particular rainfall by using the fuzzy set theory.The normalized different autocorrelation coefficient was used as weight in the moving average Markov which was used to predict rainfall state in a future period with state transition probability matrix. Then the particular rainfall was calculated according to the class designated value in fuzzy set theory. At last, the rainfall material in 30 years of the hydrological station in Tengchong and Jinghong county was used as a model. And compared with Markov model of no moving average, the error is reduced and the forecast accuracy is advanced. It shows the feasibility of the moving average-Markov model and provides a channel to be searched for advancing the prediction accuracy of the prediction of the medium-and-long-range rainfall.