Re: Regression question
- From: "Old Mac User" <chendrixstats@xxxxxxxx>
- Date: 19 Apr 2006 15:06:04 -0700
Do you mean the predicted values of the response or dependent variable
"Y" is to be constrained to values > 0 ?? If so, then build the model
in terms of Ln(Y). he model will then be of the form Ln(Y) = f(X1,
X2, etc.) Predict values of Y by exponentiating the predicted values
of Ln(Y).
For the other one, use Ln(-Y)
.
- References:
- Regression question
- From: Harry
- Regression question
- Prev by Date: Regression question
- Next by Date: Re: A methodology problem in Cox regressions
- Previous by thread: Regression question
- Index(es):
Loading