信度加权马尔科夫链预测模型及应用  

Reliability Weighted Markov Chain Prediction Model and Its Application

作　　者：夏春芳 卫彦晶 王丰效 XIA Chun-fang;WEI Yan-jing;WANG Feng-xiao(School of Mathematics and Statistics,Kashgar University,Kashgai 844000,Xinjiang,China)

机构地区：[1]喀什大学数学与统计学院,新疆喀什844000

出　　处：《兰州文理学院学报（自然科学版）》2021年第5期25-29,共5页Journal of Lanzhou University of Arts and Science(Natural Sciences)

基　　金：国家自然科学基金(10771003)。

摘　　要：利用D-S理论对原始数据建立信度函数,划分状态区间,并对不同时期历史数据进行加权,以规范化的各阶自相关系数为权重,创建加权马尔科夫链预测模型.通过特征值法对预测对象进行定量分析,计算具体预测值.根据实例进行实证对比分析,得到了较小的预测误差.Using the D-S theory to establish the reliability function of the original data and divide the state interval,a weighted Markov chain prediction model was established by weighting the historical data of different periods and taking the normalized autocorrelation coefficients of each order as weights.The prediction object was quantitatively analyzed by eigenvalue method and the specific prediction value was calculated.The empirical comparative analysis was carried out according to the example,and the smaller prediction error was obtained.

关 键 词：马尔科夫链 信度函数 加权 级别特征值 预测误差 

分 类 号：F224.9[经济管理—国民经济]