COMPAR.GEE error with logistic model

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

COMPAR.GEE error with logistic model

Charles Willis
Hello,

I am trying to run the APE program COMPAR.GEE with a model containing a
categorical response variable and a mixture of continuous and categorical
independent variables. The model runs when I have categorical (binary)
response and two continuous independent variables (VAR1 and VAR2), but when
I include a categorical (binary) independent variable (VAR3), I receive the
following output with an error:

Beginning Cgee S-function, @(#) geeformula.q 4.13 98/01/27
running glm to get initial regression estimate
  (Intercept)     VAR1 VAR2 VAR3
-2.656607e+01 -3.110687e-15 -1.582172e-16  5.313214e+01
"Error in gee(RESPONSE ~ VAR1 + VAR2 + VAR3, c(1, 1, 1, 1, 1,  :
  Cgee: error: logistic model for probability has fitted value very close to
1.
estimates diverging; iteration terminated.
In addition: Warning message:
In glm.fit(x = X, y = Y, weights = weights, start = start, etastart =
etastart,  :
  algorithm did not converge"

The input is the following model:

compar.gee(RESPONSE ~ VAR1 + VAR2 + VAR3, data = subset1, family =
"binomial", phy = prunedtree1)

I have set all of the categorical data as factors and designated the family
as "binomial". I don't know what else to do and the error message is not
clear to me. If anyone can interpret this error message and/or knows how to
run a compar.gee with a mixed set of categorical and continuous variable, I
would be greatly appreciative for your advice.

Thank you,
Charlie




--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Charles G. Willis
Department of Organismic and Evolutionary Biology
22 Divinity Ave Cambridge MA 02139
HP (857) 488-2506
WP (617) 496-3890
[hidden email]
http://www.people.fas.harvard.edu/%7Ecgwillis/
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

        [[alternative HTML version deleted]]

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Reply | Threaded
Open this post in threaded view
|

Re: COMPAR.GEE error with logistic model

Simon Blomberg-4
My guess is that this combination of variables produces separation in
the data: Too many (all?) of the response 1's are in at level of VAR3,
and the 0's are at the other level (or vice versa).

HTH,

Simon.

On Sat, 2007-12-29 at 18:39 -0500, Charles Willis wrote:

> Hello,
>
> I am trying to run the APE program COMPAR.GEE with a model containing a
> categorical response variable and a mixture of continuous and categorical
> independent variables. The model runs when I have categorical (binary)
> response and two continuous independent variables (VAR1 and VAR2), but when
> I include a categorical (binary) independent variable (VAR3), I receive the
> following output with an error:
>
> Beginning Cgee S-function, @(#) geeformula.q 4.13 98/01/27
> running glm to get initial regression estimate
>   (Intercept)     VAR1 VAR2 VAR3
> -2.656607e+01 -3.110687e-15 -1.582172e-16  5.313214e+01
> "Error in gee(RESPONSE ~ VAR1 + VAR2 + VAR3, c(1, 1, 1, 1, 1,  :
>   Cgee: error: logistic model for probability has fitted value very close to
> 1.
> estimates diverging; iteration terminated.
> In addition: Warning message:
> In glm.fit(x = X, y = Y, weights = weights, start = start, etastart =
> etastart,  :
>   algorithm did not converge"
>
> The input is the following model:
>
> compar.gee(RESPONSE ~ VAR1 + VAR2 + VAR3, data = subset1, family =
> "binomial", phy = prunedtree1)
>
> I have set all of the categorical data as factors and designated the family
> as "binomial". I don't know what else to do and the error message is not
> clear to me. If anyone can interpret this error message and/or knows how to
> run a compar.gee with a mixed set of categorical and continuous variable, I
> would be greatly appreciative for your advice.
>
> Thank you,
> Charlie
>
>
>
>
--
Simon Blomberg, BSc (Hons), PhD, MAppStat.
Lecturer and Consultant Statistician
Faculty of Biological and Chemical Sciences
The University of Queensland
St. Lucia Queensland 4072
Australia
Room 320 Goddard Building (8)
T: +61 7 3365 2506
email: S.Blomberg1_at_uq.edu.au

Policies:
1.  I will NOT analyse your data for you.
2.  Your deadline is your problem.

The combination of some data and an aching desire for
an answer does not ensure that a reasonable answer can
be extracted from a given body of data. - John Tukey.

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.