[R] How to add lines to lattice plot produced by rms::bplot
David Winsemius
dwinsemius at comcast.net
Sat Aug 14 16:33:04 CEST 2010
On Aug 14, 2010, at 9:59 AM, Frank Harrell wrote:
>
> Once you guys figure all this out, I'm glad to modify bplot to pass
> more arguments lattice if needed.
As always, Frank, I appreciate your support. In this case I think it's
not needed. What seems to be needed is simply the correct use of the
"at" argument. (I thought I had tried this before but apparently
mucked it up somehow.) This gives the desired color levels separation
and labeling of the default levelplot version of bplot output:
bldLT40 <- bplot(mdl.pred , perim=boundaries, at=c(-0.233, seq(.50,
2.25, by=0.25), seq(2.5, 5.0, by=0.5), 6:10) )
bldLT40
And this produces the expected output with its contouplot version:
bldLT40c <- bplot(mdl.pred , perim=boundaries, lfun=contourplot,
at=c(-0.233, seq(.50, 2.25, by=0.25), seq(2.5, 5.0, by=0.5), 6:10) )
bldLT40c
My only quibble with that last one was that the labels had three
digits to the right of the decimal pt but that happily went away when
I changed the low end to 0.25. All is good here.
--
David.
>
> Frank E Harrell Jr Professor and Chairman School of Medicine
> Department of Biostatistics Vanderbilt
> University
>
> On Fri, 13 Aug 2010, David Winsemius wrote:
>
>>
>> On Aug 13, 2010, at 11:25 PM, Duncan Mackay wrote:
>>
>>> Hi David
>>>
>>> I do not know if you have done something like this.
>>
>> I had tried a few efforts like that, starting with an examination of
>> str(bp.plot) as you demonstrate.
>>
>>> I tried str(bp.plot) which gave the section about the regions (for
>>> colours) as:
>>> $ panel.args.common:List of 8
>>> ..$ x : num [1:2500] 27 28 29 29.9 30.9 ...
>>> ..$ y : num [1:2500] 141 141 141 141 141 ...
>>> ..$ z : num [1:2500] -1.43 -1.41 -1.39 -1.36 -1.34 ...
>>> ..$ at : num [1:10] -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5
>>> ..$ region : logi FALSE
>>> ..$ zlab :List of 3
>>> .. ..$ label: chr "log odds"
>>> .. ..$ rot : num 90
>>> .. ..$ cex : num 1
>>> ..$ labels : logi TRUE
>>> ..$ contour: logi TRUE
>>
>> I tried (with a bplot object named "bldLT40"):
>>
>> bldLT40$legend$right$args$key$at <- c(-0.233, seq(.50, 2.25,
>> by=0.25),
>> seq(2.5, 5.0, by=0.5), 6:10)
>>
>> ... and then tried
>>
>> bldLT40$panel.args$at <- c(-0.233, seq(.50, 2.25, by=0.25), seq(2.5,
>> 5.0, by=0.5), 6:10)
>>
>> Neither of these efforts changed the boundaries beteen colors in the
>> plot area. The first effort changed the legend scal,e but that just
>> created a "misalignment" of the colors of plot area and the legend. I
>> would be interested in either a strategy that lets one alter the
>> color
>> level changes of the z variable (which in bplot-created objects is
>> "zhat", or lets one specify the values at which contour lines are
>> drawn in contourplot.
>>
>> Thanks for your efforts.
>>
>> --
>> David.
>>
>>
>>>
>>> I added the col.region and colours from a plot levelplot that I had
>>> done to see what would occur to the trellis parameters. No colours
>>> were produced when plotted.
>>> bp.plot <- bplot(p, lfun=contourplot, color.key = TRUE, col.regions
>>> =
>>> c
>>> ("#FFFFFF
>>> ","#00FFFF","#A9E2FF","#8080FF","#0000FF","#FFD18F","#FF0000") )
>>>
>>> $ panel.args.common:List of 10
>>> ..$ x : num [1:2500] 27 28 29 29.9 30.9 ...
>>> ..$ y : num [1:2500] 141 141 141 141 141 ...
>>> ..$ z : num [1:2500] -1.43 -1.41 -1.39 -1.36 -1.34 ...
>>> ..$ at : num [1:10] -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5
>>> ..$ region : logi FALSE
>>> ..$ color.key : logi TRUE
>>> ..$ zlab :List of 3
>>> .. ..$ label: chr "log odds"
>>> .. ..$ rot : num 90
>>> .. ..$ cex : num 1
>>> ..$ labels : logi TRUE
>>> ..$ contour : logi TRUE
>>> ..$ col.regions: chr [1:7] "#FFFFFF" "#00FFFF" "#A9E2FF"
>>> "#8080FF" ...
>>>
>>> So it has been added to the panel.args.common, whether you can
>>> access these are another matter.
>>>
>>> I then tried
>>> bp.plot <- bplot(p, lfun=contourplot, par.settings = list(axis.text
>>> = list(cex = 0.65)), color.key = TRUE, col.regions =
>>> c
>>> ("#FFFFFF
>>> ","#00FFFF","#A9E2FF","#8080FF","#0000FF","#FFD18F","#FF0000") )
>>>
>>> which changed the size of the axis text so it may be the case of
>>> having to add the col.regions etc to the appropriate list in
>>> par.settings
>>>
>>> I'll leave you to amend and access the colours.
>>> You may have to add values for the wireframe/levelplot arguments
>>> like "at" etc. and col.regions (I think that is the function) to
>>> produce an appropriate colour range of your choice
>>>
>>> It is a while since I have delved into these sorts of plots.
>>> Need some sustenance.
>>>
>>> Regards
>>>
>>> Duncan
>>>
>>> Duncan Mackay
>>> Department of Agronomy and Soil Science
>>> University of New England
>>> ARMIDALE NSW 2351
>>> Email home: mackay at northnet.com.au
>>>
>>>
>>>
>>> At 10:33 14/08/2010, you wrote:
>>>
>>>> I have a plot produced by function bplot (package = rms) that is
>>>> really a lattice plot (class="trellis"). It is similar to this plot
>>>> produced by a very minor modification of the first example on the
>>>> bplot help page:
>>>>
>>>> require(rms)
>>>> n <- 1000 # define sample size
>>>> set.seed(17) # so can reproduce the results
>>>> age <- rnorm(n, 50, 10)
>>>> blood.pressure <- rnorm(n, 120, 15)
>>>> cholesterol <- rnorm(n, 200, 25)
>>>> sex <- factor(sample(c('female','male'), n,TRUE))
>>>> label(age) <- 'Age' # label is in Hmisc
>>>> label(cholesterol) <- 'Total Cholesterol'
>>>> label(blood.pressure) <- 'Systolic Blood Pressure'
>>>> label(sex) <- 'Sex'
>>>> units(cholesterol) <- 'mg/dl' # uses units.default in Hmisc
>>>> units(blood.pressure) <- 'mmHg'
>>>>
>>>> # Specify population model for log odds that Y=1
>>>> L <- .4*(sex=='male') + .045*(age-50) +
>>>> (log(cholesterol - 10)-5.2)*(-2*(sex=='female') + 2*(sex=='male'))
>>>> # Simulate binary y to have Prob(y=1) = 1/[1+exp(-L)]
>>>> y <- ifelse(runif(n) < plogis(L), 1, 0)
>>>>
>>>> ddist <- datadist(age, blood.pressure, cholesterol, sex)
>>>> options(datadist='ddist')
>>>>
>>>> fit <- lrm(y ~ blood.pressure + sex * (age + rcs(cholesterol,4)),
>>>> x=TRUE, y=TRUE)
>>>> p <- Predict(fit, age, cholesterol, sex='male', np=50) # vary sex
>>>> last
>>>> bp.plot <- bplot(p, lfun=contourplot)
>>>> bp.plot
>>>>
>>>> I have tried a variety of efforts at using update (which I assume
>>>> is a
>>>> lattice function although I can find no help page for it. It does
>>>> appear in some of the lattice hep pages and my understanding is
>>>> that
>>>> it pushes objects onto the list structure of a plot object. I've
>>>> also
>>>> tried adding to it with llines()
>>>> #-------------
>>>> Oh, never mind. I recovered a memory that I had seen a solution on
>>>> rhelp and had saved it. Turns out it was from Peter Ehlers, to
>>>> whom I
>>>> offer thanks. I was trying to add a step function: ht and weight
>>>> from
>>>> a dataframe, "bld":
>>>>
>>>> trellis.focus("panel", 1, 1)
>>>> panel.lines(x=bld$inches, y=bld$BMI28, type='s')
>>>> trellis.unfocus()
>>>>
>>>> Success!
>>>>
>>>> Now... how do I control the color levels in levelplot or
>>>> contourplot???
>>>>
>>>>
>>>> David Winsemius, MD
>>>> West Hartford, CT
>>>>
>>>> ______________________________________________
>>>> 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.
>>>>
>>>
>>> ______________________________________________
>>> 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.
>>
>> David Winsemius, MD
>> West Hartford, CT
>>
>> ______________________________________________
>> 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.
>>
David Winsemius, MD
West Hartford, CT
More information about the R-help
mailing list