[R] BIC and Hosmer-Lemeshow statistic for logistic regression

[Prof Brian Ripley]

On Tue, 19 Jun 2007, spime wrote:

> Is there any windows version of Design package???

Yes, the version put up this morning works on 2.5.x (at last).
You should be able to get a Windows build now: it is showing for me on 
CRANextras (where I put it an hour or so ago: it will reach CRAN mirrors 
in a few days).

Ensure you have Hmisc 3.4-1 (on CRAN master yesterday) and Design 2.1-1.


> Frank E Harrell Jr wrote:
>>
>> spime wrote:
>>>
>>> I haven't find any helpful thread. How can i calculate BIC and 
>>> Hosmer-Lemeshow statistic for a logistic regression model. I have used 
>>> glm for logistic fit.
>>
>> See the Design package's lrm function and residuals.lrm for a better GOF
>> test.
>> --
>> Frank E Harrell Jr   Professor and Chair           School of Medicine
>>                       Department of Biostatistics   Vanderbilt University

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595