plsRcox (version 1.8.2)

coxDKplsDR: Fitting a Direct Kernel PLS model on the (Deviance) Residuals

Description

This function computes the Cox Model based on PLSR components computed model with

• as explanatory variables: a Kernel transform of Xplan.

It uses the package kernlab to compute the Kernel transforms of Xplan, then the package mixOmics to perform PLSR fit.

Usage

coxDKplsDR(Xplan, ...)
# S3 method for default
coxDKplsDR(
  Xplan,
  time,
  time2,
  event,
  type,
  origin,
  typeres = "deviance",
  collapse,
  weighted,
  scaleX = TRUE,
  scaleY = TRUE,
  ncomp = min(7, ncol(Xplan)),
  modepls = "regression",
  plot = FALSE,
  allres = FALSE,
  kernel = "rbfdot",
  hyperkernel,
  verbose = TRUE,
  ...
)
# S3 method for formula
coxDKplsDR(
  Xplan,
  time,
  time2,
  event,
  type,
  origin,
  typeres = "deviance",
  collapse,
  weighted,
  scaleX = TRUE,
  scaleY = TRUE,
  ncomp = min(7, ncol(Xplan)),
  modepls = "regression",
  plot = FALSE,
  allres = FALSE,
  dataXplan = NULL,
  subset,
  weights,
  model_frame = FALSE,
  kernel = "rbfdot",
  hyperkernel,
  verbose = TRUE,
  model_matrix = FALSE,
  contrasts.arg = NULL,
  ...
)

Value

If allres=FALSE :

cox_DKplsDR

Final Cox-model.

If allres=TRUE :

tt_DKplsDR

PLSR components.

cox_DKplsDR

Final Cox-model.

DKplsDR_mod

The PLSR model.

Details

If allres=FALSE returns only the final Cox-model. If allres=TRUE returns a list with the PLS components, the final Cox-model and the PLSR model. allres=TRUE is useful for evluating model prediction accuracy on a test sample.

References

plsRcox, Cox-Models in a high dimensional setting in R, Frederic Bertrand, Philippe Bastien, Nicolas Meyer and Myriam Maumy-Bertrand (2014). Proceedings of User2014!, Los Angeles, page 152.

Deviance residuals-based sparse PLS and sparse kernel PLS regression for censored data, Philippe Bastien, Frederic Bertrand, Nicolas Meyer and Myriam Maumy-Bertrand (2015), Bioinformatics, 31(3):397-404, doi:10.1093/bioinformatics/btu660.

See Also

coxph, plsr

Examples

data(micro.censure)
data(Xmicro.censure_compl_imp)

X_train_micro <- apply((as.matrix(Xmicro.censure_compl_imp)),FUN="as.numeric",MARGIN=2)[1:80,]
X_train_micro_df <- data.frame(X_train_micro)
Y_train_micro <- micro.censure$survyear[1:80]
C_train_micro <- micro.censure$DC[1:80]

(cox_DKplsDR_fit=coxDKplsDR(X_train_micro,Y_train_micro,C_train_micro,ncomp=6))

#Fixing sigma to compare with plsDR on Gram matrix; should be identical
(cox_DKplsDR_fit=coxDKplsDR(X_train_micro,Y_train_micro,C_train_micro,ncomp=6,
hyperkernel=list(sigma=0.01292786)))

X_train_micro_kern <- kernlab::kernelMatrix(kernlab::rbfdot(sigma=0.01292786),
scale(X_train_micro))
(cox_DKplsDR_fit2=coxplsDR(~X_train_micro_kern,Y_train_micro,C_train_micro,ncomp=6,scaleX=FALSE))

(cox_DKplsDR_fit=coxDKplsDR(X_train_micro,Y_train_micro,C_train_micro,ncomp=6,
kernel="laplacedot",hyperkernel=list(sigma=0.01292786)))

X_train_micro_kern <- kernlab::kernelMatrix(kernlab::laplacedot(sigma=0.01292786),
scale(X_train_micro))
(cox_DKplsDR_fit2=coxplsDR(~X_train_micro_kern,Y_train_micro,C_train_micro,ncomp=6,scaleX=FALSE))

(cox_DKplsDR_fit=coxDKplsDR(~X_train_micro,Y_train_micro,C_train_micro,ncomp=6))
(cox_DKplsDR_fit=coxDKplsDR(~.,Y_train_micro,C_train_micro,ncomp=6,dataXplan=X_train_micro_df))

(cox_DKplsDR_fit=coxDKplsDR(X_train_micro,Y_train_micro,C_train_micro,ncomp=6,allres=TRUE))
(cox_DKplsDR_fit=coxDKplsDR(~X_train_micro,Y_train_micro,C_train_micro,ncomp=6,allres=TRUE))
(cox_DKplsDR_fit=coxDKplsDR(~.,Y_train_micro,C_train_micro,ncomp=6,allres=TRUE,
dataXplan=X_train_micro_df))


rm(X_train_micro,Y_train_micro,C_train_micro,cox_DKplsDR_fit)