Learn R Programming

riskRegression (version 2023.12.21)

penalizedS3: S3-wrapper for S4 function penalized

Description

S3-wrapper for S4 function penalized

Usage

penalizedS3(formula, data, type = "elastic.net", lambda1, lambda2, fold, ...)

Arguments

formula

Communicated outcome and explanatory variables. See examples.

data

Data set in which formula is to be interpreted

type

String specifying the type of penalization. Should match one of the following values: "ridge", "lasso", "elastic.net".

lambda1

Lasso penalty

lambda2

ridge penalty

fold

passed to penalized::profL1

...

Arguments passed to penalized

Examples

Run this code
library(prodlim)
if (FALSE) {
## too slow
if (require("penalized",quietly=TRUE)){
library(penalized)
set.seed(8)
d <- sampleData(200,outcome="binary")
newd <- sampleData(80,outcome="binary")
fitridge <- penalizedS3(Y~X1+X2+pen(7:8), data=d, type="ridge",
                standardize=TRUE, model="logistic",trace=FALSE)
fitlasso <- penalizedS3(Y~X1+X2+pen(7:8), data=d, type="lasso",
                standardize=TRUE, model="logistic",trace=FALSE)
# fitnet <- penalizedS3(Y~X1+X2+pen(7:8), data=d, type="elastic.net",
# standardize=TRUE, model="logistic",trace=FALSE)
predictRisk(fitridge,newdata=newd)
predictRisk(fitlasso,newdata=newd)
# predictRisk(fitnet,newdata=newd)
Score(list(fitridge),data=newd,formula=Y~1)
Score(list(fitridge),data=newd,formula=Y~1,split.method="bootcv",B=2)
data(nki70) ## S4 fit
fitS4 <- penalized(Surv(time, event), penalized = nki70[,8:77],
                 unpenalized = ~ER+Age+Diam+N+Grade, data = nki70,
                 lambda1 = 1)
fitS3 <- penalizedS3(Surv(time,event)~ER+Age+Diam+pen(8:77)+N+Grade,
                     data=nki70, lambda1=1)
## or
penS3 <- penalizedS3(Surv(time,event)~ER+pen(TSPYL5,Contig63649_RC)+pen(10:77)+N+Grade,
                     data=nki70, lambda1=1)
## also this works
penS3 <- penalizedS3(Surv(time,event)~ER+Age+pen(8:33)+Diam+pen(34:77)+N+Grade,
                    data=nki70, lambda1=1)
}}

Run the code above in your browser using DataLab