predict.big_pls_cox_gd

Provides Partial least squares Regression and various regular, sparse
or kernel, techniques for fitting Cox models for big data. Provides a Partial
Least Squares (PLS) algorithm adapted to Cox proportional hazards models that
works with 'bigmemory' matrices without loading the entire dataset in memory.
Also implements a gradient-descent based solver for Cox proportional hazards
models that works directly on 'bigmemory' matrices.
Bertrand and Maumy (2023) <https://hal.science/hal-05352069>, and
<https://hal.science/hal-05352061> highlighted
fitting and cross-validating PLS-based Cox models to censored big data.

Frederic Bertrand

bigPLScox

Partial Least Squares for Cox Models with Big Matrices

Myriam Maumy-Bertrand

predict.big_pls_cox_gd function

<dl><dt>object</dt>
<dd>A big_pls_cox_gd model.</dd>
<dt>newdata</dt>
<dd>A numeric matrix or big.matrix. If NULL, uses training scores.</dd>
<dt>type</dt>
<dd>One of "link", "risk", "response", "components".</dd>
<dt>comps</dt>
<dd>Integer vector of components to use. Default: all.</dd>
<dt>coef</dt>
<dd>Optional coefficient vector. If NULL, uses coef(object$cox_fit).</dd>
<dt>...</dt>
<dd>Unused.</dd></dl>

Arguments

Predict method for big_pls_cox_gd — predict.big_pls_cox_gd

<dl>

<dt>object</dt>
<dd>A big_pls_cox_gd model.</dd>


<dt>newdata</dt>
<dd>A numeric matrix or big.matrix. If NULL, uses training scores.</dd>


<dt>type</dt>
<dd>One of "link", "risk", "response", "components".</dd>


<dt>comps</dt>
<dd>Integer vector of components to use. Default: all.</dd>


<dt>coef</dt>
<dd>Optional coefficient vector. If NULL, uses coef(object$cox_fit).</dd>


<dt>...</dt>
<dd>Unused.</dd>

</dl>

predict.big_pls_cox_gd: Predict method for big_pls_cox_gd

Description

Usage

Arguments