Fitted Value Too Large
Posted: Mon Apr 30, 2012 5:46 pm
by stoddj
This seems silly, but my fitted LHS value is way too large. I'm hitting the Forecast button after doing an OLS, in order to generate a fitted version of the LHS variable. But that fitted value is way too large -- by a factor of 100! I can otherwise generate it by just subtracting the regression's residuals from the actual value of the original LHS variable. But this is an extra step. Any idea what I'm doing wrong? I've re-read the instructions, and nary a clue do I find. Thanks, JS
Fitted Value Too Large
Posted: Tue May 01, 2012 12:01 am
by EViews Gareth
You should provide more details such as the equation specification and post the workfile.
Re: Fitted Value Too Large
Posted: Tue May 01, 2012 6:29 am
by stoddj
Great -- now it's not doing it anymore. I guess only a magician can diagnose a problem that's not happening.
Here's the file, for what it's worth. It should be available at
http://www.ewp.rpi.edu/hartford/~stoddj/oldkauf.wf1
The tab is OLS, and the equation is Lifex_OLS. The (now correct) fitted variable is Lifexf. I scrubbed the old bad one, unfortunately.
Thanks anyway! // JS