xprcomp: compute principal projection simulating out of sample data.

In WinVector/crosspca: Cross Validated Principal Components Analysis

Description

compute principal projection simulating out of sample data.

Usage

xprcomp(x, k, crossplan, center, scale.)

Arguments

x	matrix to perfrom principal components analysis on.
k	integer number of components to estimate.
crossplan	list of lists which have train and app slots, where app is a partition of seq_len(nrow(x)).
center	logical controls centering of data.
scale.	logical controls scaling of data.

Value

training (simulated out of sample projected) matrix and projection.

Examples

set.seed(23552)
splitPlan <- vtreat::kWayCrossValidation(nrow(USArrests),3,NULL,NULL)
dc <- xprcomp(USArrests,k=2,crossplan=splitPlan,center=TRUE,scale.=TRUE)

WinVector/crosspca documentation built on June 24, 2020, 5:09 p.m.