基于皮尔逊层次聚类的马尔可夫优化模型在降水预测中的应用

Application of Markov optimization model based on pearson hierarchical clustering in precipitation prediction

以黄委青铜峡水文站年降水数据为例,选取Kolmogorov-Smirnov分布检验对现有青铜峡水文站年降水数据进行降水分布显著性检验,判断历史资料的分布模型。采用基于皮尔逊层次聚类法进行状态划分,建立最优预测模型,并根据已确定的降水分布,通过K-S检验对该地区未来降水预测的准确性进行检验。结果表明,青铜峡水文站降水分布为皮尔逊三型分布;马尔可夫模型能够适用于青铜峡地区降水预测。

Taking the annual precipitation data of Qingtongxia hydrological station as an example,the Kolmogorov Smirnov distribution test is selected to test the significance of precipitation distribution of the annual precipitation data of Qingtongxia hydrological station,and judge the distribution model of historical data.The hierarchical superposition method based on Pearson is used to divide the States,establish the optimal prediction model,verify the determined precipitation distribution according to the existing data,and test the accuracy of future precipitation prediction in the region by K-S test.The results show that the precipitation distribution of Qingtongxia hydrologic station is Pearson three type distribution;The Markov model can be applied to the precipitation prediction in Qingtongxia area.