recover_pca_data: Recover data from principal component analysis

In doBy: Groupwise Statistics, LSmeans, Linear Estimates, Utilities

Recover data from principal component analysis

Description

Recover data from principal component analysis based on the first (typically few) components.

Usage

recover_pca_data(object, comp = 1)

Arguments

object	An object of class prcomp.
comp	The number of components to be used. Must be smaller than the number of variables.

Value

A dataframe

Examples

crime <- doBy::crimeRate
rownames(crime) <- crime$state
crime$state <- NULL

o <- order(apply(scale(crime), 1, sum))
dat <- crime[o,]
head(dat)
tail(dat)
matplot(scale(dat), type="l")

pc1 <- prcomp(dat, scale. = TRUE)
summary(pc1)
rec2 <- recover_pca_data(pc1, 2)

pairs(rec2)

par(mfrow=c(1,2))
matplot(scale(dat), type="l")
matplot(scale(rec2), type="l")

j <- merge(dat, rec2, by=0)
pairs(j[,-1])

doBy documentation built on Oct. 8, 2024, 1:06 a.m.