R-Packages (Statistical Computing)

Comment

Comment by Alex Guazzelli on October 13, 2008 at 11:37am

Hi all! As you many know, PMML is a standard XML-based language used to represent predictive models. In this way, models built with your favorite statistical tool (R) can be exported into PMML and imported into another tool. The good news is that a PMML package is now available for R (you can find the latest in CRAN).

Anyway, I just created a PMML group here at AnalyticBridge and would like to invite you to join us. I hope to have a good amount of productive debate on PMML as well as different ways packages could support it.

Thanks,

Alex

Comment by John Logsdon on September 24, 2008 at 7:16am

It works OK - try for example:

summary(survreg(Surv((y<-rnorm(100,10,1))-0.1,y+0.1,rep(3,100),"interval")~1,dist="gaussian")

ie just generate y = 100 rnorms, mean 10, sd 1 and do an interval fit for data between y-0.1 and y+0.1. GIves more or less the same answer as summary(lm(y~1))

Comment by Muhammad Riaz on August 26, 2008 at 8:39am

Dear Pryseley,
Thank you for your query in this group, It is a bit late to reply, however it may help if you still need help for your problem.
I think you can fit parametric survival models, if you want to use R- then please see R help relating to the command line given below:

survreg(formula=formula(data), data=parent.frame(), weights,
subset,na.action,dist="weibull", init=NULL, scale=0,
control=survreg.control(),parms=NULL,model=FALSE, x=FALSE,
y=TRUE, robust=FALSE, ...)

However it is very easy in Stata to use the command given below for fitting a parametric survival model. Again please see help in stata for streg.
streg

Comment by Pryseley Nkouibert on June 16, 2008 at 9:39am

I wish to fit parametric survival models for interval censored data. However, I run into trouble while using the survreg function.
I used survival object of the form Surv(Start, Stop, Event) and I get a warning message indicating "invalid survival type". Is it not
possible to fit parametric survival models using the survreg function? If so, is there any package in R which enables fitting of such
models?

Best Regards,
Pryseley