[R] coefficient of logistic regression

[orkun]

John Fox wrote:

> Dear Ahmet,
>
> Sorry for the slow response, but I've been busy all today, 
> coincidentally teaching a workshop on logistic regression.
>
> Tom Blackwell sent you a useful suggestion for interpreting 
> coefficients on the odds scale. If you want to trace out the partial 
> relationship of the fitted probability of response to a particular 
> predictor holding others constant, you can set the other predictors to 
> typical values and let the predictor in question vary over its range, 
> transforming the fitted log-odds to the probability scale.
>
> You may be interested in my effects package (on CRAN or at 
> <http://socserv.socsci.mcmaster.ca/jfox/Misc/effects/index.html>), 
> which makes these kinds of displays for linear and generalized-linear 
> models, including those with interactions.
>
> Regards,
>  John
>
> At 03:06 PM 6/3/2003 +0300, orkun wrote:
>
>> John Fox wrote:
>>
>>> At 11:54 AM 6/3/2003 +0300, orkun wrote:
>>>
>>>> in logistic regression,
>>>> I want to know that it is possible to get probability values of 
>>>> each predictors by
>>>> using following formula for each predictor one by one (keeping 
>>>> constant the others)
>>>> <<< exp(coef)/(1+exp(coef)) >>>
>>>
>>>
>>>
>>> Dear Ahmet,
>>>
>>> This will almost surely give you nonsense, since it produces a 
>>> fitted probability ignoring the constant in the model (assuming that 
>>> there is one), setting other predictors to 0 and the predictor in 
>>> question to 1. What is it that you want to do?
>>>
>>> I hope that this helps,
>>>  John
>>>
>>>
>> thank you
>>
>> Say, I just want to find each predictor's particular effect on 
>> dependent variables.
>> Actual model is to prepare landslide susceptibility map on GIS. So  I 
>> want to know
>> what the effect as probability value comes from each predictor. For 
>> instane what is the effect
>> of  slope on landslide susceptibility. Should I keep others constant ?
>>
>> kind regards
>
>
>
>
>
> -----------------------------------------------------
> John Fox
> Department of Sociology
> McMaster University
> Hamilton, Ontario, Canada L8S 4M4
> email: jfox at mcmaster.ca
> phone: 905-525-9140x23604
> web: www.socsci.mcmaster.ca/jfox
> -----------------------------------------------------
>
>
Dear Mr. Fox

thank you very much all.

Because of related to your answer. I ask you directly if you don't mind
I studied several ways after my email.
I wonder whether pgeo<-predict.glm(glm.ob,type="terms")
gives same result with probability value I asked before.
I tried on it. But it gives "Error in rep(1/n,n) %*% 
model.matrix(object): non conformable
arguments" .

By the way , your teaching notes is available on the internet ?


cordially

can y



______________________________________



______________________________________
The views and opinions expressed in this e-mail message are the ... {{dropped}}