Dear R users, I picked up ARIMA(2,3,0) model for my time series analysis using auto.arima() function.
>From my understanding, the model should be in this shape: [Y(hat)(t)-Y(t-1)]-2[Y(t-1)-Y(t-2)]+[Y(t-2)-Y(t-3)]=Theta(1){[(Y(t-1)-Y(t-2)]-2[(Y(t-2)-Y(t-3)]+[Y(t-3)-Y(t-4)]}+Theta(2){[Y(t-2)-Y(t-3)]-2[Y(t-3)-Y(t-4)]+[Y(t-4)-Y(t-5)]}+mu(e(t)) which in this case, mu(e(t))=0, cause the residuals' mean is assumed and estimated as zero. In R, when t is 2013, the estimation value of Y(hat) is 7.084688, which is different from my own calculation 7.1882, according to the model above. Do you think that I missed something? And is there any way that I can simplify the model or the differential equation above??? Many thanks, Chintemur Batur ############################################################################ Here is the code library(tseries) library(forecast) d=scan("D:/Data.txt") d D=ts(data=d, start=1981,end=2012, frequency=1) D ############################################################################## Time Series: Start = 1981 End = 2012 Frequency = 1 [1] 384 403 427 450 499 550 575 615 640 680 702 730 760 790 [15] 790 830 870 871 906 920 968 1010 1060 1111 1165 1191 1217 1221 [29] 1089 1089 1090 1103 ############################################################################# lnD=log(D) lnD3=diff(lnD, differences=3) adf.test(lnD3) par(mfrow=c(3,1)) acf(lnD3, lag.max=20) pacf(lnD3, lag.max=20) autoarima=auto.arima(lnD,d=3) summary(autoarima) Box.test(autoarima$residuals,lag=20,type="Ljung-Box") ForecastAutoArima=forecast.Arima(autoarima, h=5, c=(0.95)) plot.forecast(ForecastAutoArima) [[alternative HTML version deleted]] ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.