The Realization Of The Optimal Partition Method Algorithm And Its Application

The optimal partition method [OPM], is a kind of method of the ordered clustering, which can be used to classify the annual rainfall for some area in some period. Because it needs much complicated computation to compute the variation, it is not easy for operation in reality. Many authors have done theoretical and methodical analysis. In this paper, the Matlab program language is offered to carry out the algorithm for OPM.Firstly, the predication of weighed Markov chain is to set up the classified standard. Secondly, adopt all kinds of the standardized self-coefficient as weights. Finally, predict the trend of rainfall in virtue of those weights. So the classified standard deeply affects the result of weighed predication. For this reason, the paper discusses the different impacts that OPM and the sample-mean variance classified method have made on the precision of predicting the rainfall of weighed Markov chain by the technology of stochastic simulation.With the increase of data, the computation for weighed predication becomes very complicated. For convenience, the paper also makes a program by Matlab language to realize the algorithm, which offers the technical guarantee for stochastic simulation.During discussion, the paper sets the data of annual rainfall of the city of Wuhan from 1951 to 2004 as an example and has a good analysis of substantial evidence for weighed Markov chain predication. What's more, the paper predicts the rainfall of 2008. The result implies that it is semiarid year and the annual rainfall varies between 968mm and 1184mm. So the city of Wuhan should make a preparation for combating drought.Finally, the paper obtains the index of flood by the annual rainfall of the city of Wuhan. The author is convinced of the index of flood is helpful for combating flood and drought of the city of Wuhan in the future.