[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: fmlogit fit assessment & predictive ability?

From   Leslie Witter <[email protected]>
To   [email protected]
Subject   st: fmlogit fit assessment & predictive ability?
Date   Sun, 31 Jan 2010 11:38:06 -0800 (PST)


I'm using fmlogit to explore factors affecting proportion of time spent in various activities.  Two questions about postestimation:

First, I'm getting a trend in residuals where the model underpredicts the relatively rare cases when one activity dominates all/most of an observation period.  This is consistent across all activities.  Has anyone had this type of issue with fmlogit; any ideas of how I could get a better fit?

Second, do you know of a way other than a residual plot to get at fit assessment/predictive ability for fmlogits?

Thanks so much!


Leslie Witter

M.Sc. candidate - NRES Biology

University of Northern British Columbia


*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index