| Type: | Package |
| Imports: | MASS, mvtnorm, glmnet, splines, survival, cvTools |
| Depends: | foreach, parallel |
| Title: | Broken Adaptive Ridge AFT Model with Censored Data |
| Version: | 0.1.1 |
| Description: | Broken adaptive ridge estimator for censored data is used to select variables and estimate their coefficients in the semi-parametric accelerated failure time model for right-censored survival data. |
| License: | GPL-2 |
| RoxygenNote: | 7.0.2 |
| NeedsCompilation: | no |
| Packaged: | 2020-11-30 12:31:55 UTC; zhihu |
| Author: | Zhihua Sun [aut, cre], Chunyu Yu [aut], Gang Li [aut], Kani Chen [ctb], Yi Liu [ctb] |
| Maintainer: | Zhihua Sun <zhihuasun@ouc.edu.cn> |
| Repository: | CRAN |
| Date/Publication: | 2020-11-30 13:10:08 UTC |
Broken Adaptive Ridge Estimator for Censored Data in AFT Model
Description
Prints 'Broken adaptive ridge (BAR) method to the semi-parametric accelerated failure time (AFT) model for right-censored survival data by applying the Leurgan's synthetic data.'.
Usage
CenBAR(X,Y,delta,lambda.path=NULL, enableScreening=FALSE)
Arguments
X |
input matrix, of dimension nobs x nvars; each row is an observation vector. |
Y |
response variable. |
delta |
The status indicator, normally 0=alive, 1=dead. |
lambda.path |
A user supplied lambda sequence. One usage is to have the program compute its own lambda sequence based on nlambda and lambdaMax. lamdMax = max((t(x)*Y)^2/(4*t(x)*x)). The other usage is use the sequence depend on user's data. |
enableScreening |
If nobs > nvars, there is no need to do screening; If nobs <= nvars, it will do variable screening and then variable selection and estimate (defalt is FALSE). |
Value
beta |
the coefficients estimation of the variables. |
Author(s)
Zhihua Sun, Chunyu Yu
Examples
X=matrix(rnorm(10*2),10,2)
Y=abs(rnorm(10))
delta=sample(0:1,10,replace=TRUE)
lambda.path <- seq(0.1, 10, l=5)
fit = CenBAR(X,Y,delta,lambda.path)