R help - Nov 2010 - calculating martingale residual on new data using "predict.coxph"

Hi list,

I was trying to use "predict.coxph" to calculate martingale residuals
on a test
data, however, as pointed out before

http://tolstoy.newcastle.edu.au/R/e4/help/08/06/13508.html

predict(mycox1, newdata, type="expected") is not implemented yet. 
Dieter
suggested to use 'cph' and 'predict.Design', but from my reading
so far, I'm not
sure they can do that.

Do you other ways to calculate martingale residuals on a new data?

Thank you very much!

...Tao

On Nov 19, 2010, at 12:32 PM, Shi, Tao wrote:
> Hi list,
>
> I was trying to use "predict.coxph" to calculate martingale  
> residuals on a test
> data, however, as pointed out before
What about resid(fit) ?  It's my reading of Therneau & Gramsch [and of  
help(coxph.object) ] that they consider those martingale residuals.
>
> http://tolstoy.newcastle.edu.au/R/e4/help/08/06/13508.html
>
> predict(mycox1, newdata, type="expected") is not implemented yet.
> Dieter
> suggested to use 'cph' and 'predict.Design', but from my
reading so
> far, I'm not
> sure they can do that.
>
> Do you other ways to calculate martingale residuals on a new data?
>
> Thank you very much!
>
> ...Tao

David Winsemius, MD
West Hartford, CT

This feature has been added in survival 2.36-1, which is now on CRAN.
(2.36-2 should appear in another day or so)
     Terry T.

---------begin included message --------
I was trying to use "predict.coxph" to calculate martingale residuals
on
a test 
data, however, as pointed out before

http://tolstoy.newcastle.edu.au/R/e4/help/08/06/13508.html

predict(mycox1, newdata, type="expected") is not implemented yet.

Thank you, Terry!




----- Original Message ----
> From: Terry Therneau <therneau at mayo.edu>
> To: "Shi, Tao" <shidaxia at yahoo.com>
> Cc: r-help at r-project.org; dieter.menne at menne-biomed.de; r_tingley at
hotmail.com
> Sent: Mon, November 22, 2010 6:11:15 AM
> Subject: Re:  calculating martingale residual on new data using 
"predict.coxph"
> 
> This feature has been added in survival 2.36-1, which is now on CRAN.
> (2.36-2  should appear in another day or so)
>      Terry  T.
> 
> ---------begin included message --------
> I was trying to use  "predict.coxph" to calculate martingale
residuals on
> a test 
> data,  however, as pointed out  before
> 
> http://tolstoy.newcastle.edu.au/R/e4/help/08/06/13508.html
> 
> predict(mycox1,  newdata, type="expected") is not implemented
yet.
> 
>

Thank you for the advice, Frank!

...Tao



----- Original Message ----
> From: Frank Harrell <f.harrell at vanderbilt.edu>
> To: r-help at r-project.org
> Sent: Sun, November 21, 2010 5:49:36 AM
> Subject: Re: [R] calculating martingale residual on new data using
> 
> 
> The tendency is to use residual-like diagnostics on the entire dataset 
that
> was available for model development.  For test data we typically  run
> predictive accuracy analyses.  For example, one of the strongest 
validations
> is to show, in a high-resolution calibration plot, that absolute 
predictions
> (e.g., probability of survival at 2 years) are  accurate.
> 
> Frank
> 
> 
> -----
> Frank Harrell
> Department of  Biostatistics, Vanderbilt University
> -- 
> View this message in context:  
>http://r.789695.n4.nabble.com/calculating-martingale-residual-on-new-data-using-predict-coxph-tp3050712p3052377.html
>
> Sent  from the R help mailing list archive at Nabble.com.
> 
> ______________________________________________
> R-help at r-project.org 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.
>