基于加权马尔可夫链预测方法最优分组的实证分析研究  

An Empirical Analysis of Classification Standards Based on the Weighted Markov Chain Prediction Method

作　　者：张杰[1] 夏乐天[1] 

机构地区：[1]河海大学理学院数学系,江苏南京210098

出　　处：《数学的实践与认识》2015年第6期172-179,共8页Mathematics in Practice and Theory

摘　　要：在介绍加权马尔可夫链预测方法的基础上,以昆明市东川区1956年至2000年的年降雨量为实例,分别基于样本均值-均方差分组法和有序聚类分组法(最优分割法)进行加权马尔可夫链滚动预测的实证分析.结果表明有序聚类分组法比均值-均方差分组法更为合理.Based on the annual precipitation data from 1956 to 2000 Dong-Chuan District,the weighted Markov chain is applied to prediction and analysis of states of the rainfall,in which the sample mean-standard deviation classified method and optimum partition method are taken as the classification standard respectively.It is concluded that the optimum partition method is more reasonable than the mean-standard deviation classified method.

关 键 词：加权马尔可夫链 预测 最优分割法 降雨量 

分 类 号：O211.62[理学—概率论与数理统计]