In the context of predictive modeling, one builds a model on
a training sample and then assesses its accuracy via a validation model.
Reporting the r-squared, mean absolute prediction error implied in the
validation sample is often necessary. Is there an automated way to get these
out-of-sample results canned in Stata or available in an ado file? I usually
manually code these. Thanks everyone.
Chris
*******************************
M. Christopher Roebuck
Health Economist / Manager
Predictive Modeling & Forecasting
Research & Development
(410) 785-2136 Direct
(410) 785-8140 Fax
[email protected]
Caremark
One Hunt
Valley
11311 McCormick
Road
Suite 230
Hunt Valley, MD 21031
*******************************
