【优质】以选择简单模式解释长期走势为原则.pptVIP

* Chap6 Time Series Regression Modeling trend : polynomial functions Detecting Autocrrelation : Residual plot Durbin-Watson statistics Handling First-order autocorrelation Modeling seasonal variation: dummy variable Growth curve model 6.1 Modeling trend Yt = TRt + εt , Trend : linear trend, curvilinear trend Model with polynomial trend Yt = TRt + εt = β0+ β1 t + β2 t2 …….+ βp t p + εt , εt ~NID(0,σ2) Model with power trend Yt = β0( β1 t ) εt , log(Yt) = logβ0+ log(β1 ) t +logεt logεt ~NID(0,σ2) curvilinear