数学建模案例与方法教学课件第9章统计数据分析方法.pptVIP

My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template My First Template 9.5 时间序列分析方法 模型的识别 9.5.1 9.5 时间序列分析方法 自回归AR(p)模型的判断 1. p阶自回归模型AR（p）的形式为 Yn=φ1Yn-1+…+φpYn-p+εn（9-18） 式中，φ1,…,φp为未知参数；Yn为因变量，即它可由以前各时间点数值表示；εn为白噪声序列，其数学期望为零，方差为σ2ε，它代表了不能用模型说明的随机因素。 对给定的k，计算φ^ii,ik；观察满足φ^ii≤2/n≤2/n的φ^ii比例是否达到了95.5%。若在 k=1,2,…,p-1时均没有达到，而在k=p时达到了，则称φ^kk在p以后截尾。这说明此序列为AR(p)序列。 9.5