R help - Aug 2016 - Labelling interaction axes of a nomogram using rms

Good afternoon,

I have tried searching and have hit a wall with my own trial and error. I
have been able to use "nice" labels for non-interacted variables by
setting
them using Newlabels(), but I am unable to do so for the interaction terms
following a fit.

Can you please offer a suggestion?

Kind regards,
Leonardo

	[[alternative HTML version deleted]]

> On Aug 26, 2016, at 11:53 AM, Leonardo Guizzetti <leonardo.guizzetti at
gmail.com> wrote:
> 
> Good afternoon,
> 
> I have tried searching and have hit a wall with my own trial and error. I
> have been able to use "nice" labels for non-interacted variables
by setting
> them using Newlabels(), but I am unable to do so for the interaction terms
> following a fit.
> 
> Can you please offer a suggestion?
My suggestion is that you post a minimal example.

-- 
David.
> 
> Kind regards,
> Leonardo
> 
> 	[[alternative HTML version deleted]]
> 
> ______________________________________________
> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> 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.
David Winsemius
Alameda, CA, USA

Hi David,

As suggested, here is some example code adapted from the example of the
nomogram() from the rms package.  You'll notice that variable labels are
not being applied to main effects or interaction terms.

Leonardo

rm(list = ls())
# Import libraries
require(Hmisc)
require(rms)

set.seed(12345)
n <- 1000 # define sample size
age <- rnorm(n, 0, 10)
sbp <- rnorm(n, 120, 15)
sex <- factor(sample(c(0,1), n, TRUE), levels=c(0,1),
labels=c('female','male'))
# Specify population model for log odds that Y=1
L <- 0.5*(sex=="male") + 0.01*age + 0.12*sbp + 0.02*sbp*age +
rnorm(n, 0,
10)
# Simulate binary y to have Prob(y=1) = 1/[1+exp(-L)]
y <- ifelse(runif(n) < plogis(L), 1, 0)

# Create a dataframe and apply nice labels to each variable.
mydata <- data.frame(age, sbp, sex, y)
var.labels <- c(age = "Patient Age - 50 years",
                sbp = "Systolic blood pressure",
                sex = "Sex",
                y = "Outcome")
label(mydata) = lapply(names(var.labels),
                       function(x) label(mydata[,x]) <- var.labels[x])
label(mydata) # show the labels.

# Fit a model
ddist <- datadist(mydata)
options(datadist="ddist")
f <- rms::lrm(y ~ age*sbp + sex)
f
f$coefficients

# Create a nomogram.
nom <- nomogram(f, fun=plogis, lp=F, nint=10, maxscale=100,
                interact = list( sbp=c(100,140,180) ),
                fun.at=c(.01,.05,seq(.1,.9,by=.1),.95,.99),
                funlabel="Risk of Death", vnames="labels")
plot(nom)

# end
options(datadist=NULL)


On Fri, Aug 26, 2016 at 10:35 PM, David Winsemius <dwinsemius at
comcast.net>
wrote:
>
> > On Aug 26, 2016, at 11:53 AM, Leonardo Guizzetti <
> leonardo.guizzetti at gmail.com> wrote:
> >
> > Good afternoon,
> >
> > I have tried searching and have hit a wall with my own trial and
error. I
> > have been able to use "nice" labels for non-interacted
variables by
> setting
> > them using Newlabels(), but I am unable to do so for the interaction
> terms
> > following a fit.
> >
> > Can you please offer a suggestion?
>
> My suggestion is that you post a minimal example.
>
> --
> David.
> >
> > Kind regards,
> > Leonardo
> >
> >       [[alternative HTML version deleted]]
> >
> > ______________________________________________
> > R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> > 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.
>
> David Winsemius
> Alameda, CA, USA
>
>
	[[alternative HTML version deleted]]