Hi, I have been using this website ( http://www.stat.pitt.edu/stoffer/tsa2/Rissues.htm http://www.stat.pitt.edu/stoffer/tsa2/Rissues.htm ) to help me to fit ARIMA models to my data. At the moment I have two possible methods to use. Method 1 If I use arima(ts.data, order=c(1,2,0), xreg=1:length(ts.data)) then the wrong value for the intercept/mean is given (checked on SPSS and Minitab) and also, this is produced In sqrt(diag(x$var.coef)) : NaNs produced Which means that the t-values (for the coefficients) are NaNs, which in turn means that the p-values are NaNs. Although, using this method gives the correct forecast (using predict) and enables ts.plot to show the forecast and 95% CI's. Method 2 If I use diff(diff(ts.dat)) and then apply an ARIMA(1,0,0) to it, then this gives the correct coefficients but the forecasts are wrong (ie they are flat and do not follow the trend). Could anyone think of a way to get both the coefficients AND the forecasts correct? Thanks. -- View this message in context: http://www.nabble.com/Time-Series---ARIMA-differencing-problem-tp22353903p22353903.html Sent from the R help mailing list archive at Nabble.com.
Hi, I have been using this website ( http://www.stat.pitt.edu/stoffer/tsa2/Rissues.htm http://www.stat.pitt.edu/stoffer/tsa2/Rissues.htm ) to help me to fit ARIMA models to my data. At the moment I have two possible methods to use. Method 1 If I use arima(ts.data, order=c(1,2,0), xreg=1:length(ts.data)) then the wrong value for the intercept/mean is given (checked on SPSS and Minitab) and also, this is produced In sqrt(diag(x$var.coef)) : NaNs produced Which means that the t-values (for the coefficients) are NaNs, which in turn means that the p-values are NaNs. Although, using this method gives the correct forecast (using predict) and enables ts.plot to show the forecast and 95% CI's. Method 2 If I use diff(diff(ts.dat)) and then apply an ARIMA(1,0,0) to it, then this gives the correct coefficients but the forecasts are wrong (ie they are flat and do not follow the trend). Could anyone think of a way to get both the coefficients AND the forecasts correct? Thanks. -- View this message in context: http://www.nabble.com/Time-Series---ARIMA-differencing-problem-tp22354071p22354071.html Sent from the R help mailing list archive at Nabble.com.
Uploaded the data and my comparison. Hopefully this will help illustrate and solve the problem. http://www.nabble.com/file/p22371555/data.csv data.csv http://www.nabble.com/file/p22371555/arima%2Bmethods.docx arima+methods.docx -- View this message in context: http://www.nabble.com/Time-Series---ARIMA-differencing-problem-tp22354071p22371555.html Sent from the R help mailing list archive at Nabble.com.
The docx file (which includes graphs) can be opened with either Microsoft Office 2007 or if you have an earlier version: Google for the compatibility pack ...OR use openoffice 3. thefurryblur wrote:> > Uploaded the data and my comparison. Hopefully this will help illustrate > and solve the problem. > http://www.nabble.com/file/p22371555/data.csv data.csv > http://www.nabble.com/file/p22371555/arima%2Bmethods.docx > arima+methods.docx >-- View this message in context: http://www.nabble.com/Time-Series---ARIMA-differencing-problem-tp22354071p22372784.html Sent from the R help mailing list archive at Nabble.com.