Which version of Stata are you using? I ask because, if you are using
Stata Version 8, 7 or 6, then you will not be able to run the Stata 9
version of -somersd-, which can be downloaded from SSC. Instead, you must
download the latest Stata 6 version of -somersd- from my website by
typing, in Stata,
net from http://www.kcl-phs.org.uk/rogernewson/stata6/
and then following the instructions for downloading the Stata 6 -somersd-
package. Alternatively, you can upgrade to Stata 9 and then download the
Stata 9 version of -somersd- from SSC.
Neither the Stata 6 version of -somersd- nor the Stata 9 version
of -somersd- have any way of knowing whether or not the predictor variable
is a score from a logistic regression model. However, to define a logistic
regression score, you must use the -predict- command, usually with
the -xb- option.
I hope this helps.
Best wishes
Roger
At 21:24 21/11/2005, Luo wrote:
Dear Roger,
I have a couple of questions about somersd. First, when I tried to run
somersd, I got an error message "error occurred while loading
somersd.ado". I assume the error had to do with the locations of the
related files. I have STAT under c:\STAT8, but the somersd files seemed
to be automatically downloaded to c:\DATA and c:\ado. Would these file
locations cause any problems? My second question is directly about
somersd commands. Do I run the somersd commands after I run the logistic
commands? How does somersd know that I am computing Harrell's C with
regard to a particular set of dependent and independent variables for a
logistic regression model? I am new with STATA, so please bear with me if
my questions are somehow elementary. Thanks in advance for any helps from
you!
Luo
----- Original Message ----- From: "Roger Newson" <[email protected]>
To: <[email protected]>
Sent: Monday, 21 November 2005 7:34 AM
Subject: Re: st: Jackknife Cross-Validation for Logistic Regression Models
At 02:57 21/11/2005, Luo wrote:
I am trying to use STATA to do Jackknife cross-validations of the
classification results from logistic regression models. The Jackknife
cross-validations I have in mind include the classification tables and
ROC curves. I know STATA has a JKnife procedure applicable to various
statistical analyses, including logistic regression models, but can I
make the procedure produce specifically the above kinds of outcomes? Or
are there any other alternatives for these outcomes? Thanks!
Jackknife confidence intervals for the ROC area can be produced using
the -somersd- package, which you can download from SSC using the -ssc-
command. The -somersd- package includes a .pdf manual -somersd.pdf-,
which discusses methods for estimating Harrell's C (equivalent to the ROC
area) for models estimated from the same dataset.
I hope this helps.
Roger
--
Roger Newson
Lecturer in Medical Statistics
Department of Public Health Sciences
Division of Asthma, Allergy and Lung Biology
King's College London
5th Floor, Capital House
42 Weston Street
London SE1 3QD
United Kingdom
Tel: 020 7848 6648 International +44 20 7848 6648
Fax: 020 7848 6620 International +44 20 7848 6620
or 020 7848 6605 International +44 20 7848 6605
Email: [email protected]
Website: http://phs.kcl.ac.uk/rogernewson/
Opinions expressed are those of the author, not the institution.
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
--
Roger Newson
Lecturer in Medical Statistics
Department of Public Health Sciences
Division of Asthma, Allergy and Lung Biology
King's College London
5th Floor, Capital House
42 Weston Street
London SE1 3QD
United Kingdom
Tel: 020 7848 6648 International +44 20 7848 6648
Fax: 020 7848 6620 International +44 20 7848 6620
or 020 7848 6605 International +44 20 7848 6605
Email: [email protected]
Website: http://phs.kcl.ac.uk/rogernewson/
Opinions expressed are those of the author, not the institution.
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
