lixiaoyao5880@YAHOO.COM wrote back:
>yes,this data is for weeks(1-508 weeks),I used the time series to solve
>this question and get pretty good result. Do you have new comments on
Did you *really* use time series method to do this? It looked to me like
just threw some lags into PROC REG and ignored the collinearity issues. But
you never showed us any of your work, other than some of your
METHOD=RSQUARE results, so I cannot give you any decent advice.
David L. Cassell
3115 NW Norwood Pl.
Corvallis OR 97330
Express yourself instantly with MSN Messenger! Download today - it's FREE!