prcompgdsn: Principal Component Analysis for high-dimensional data

In TJGorrie/bigmelon: Illumina methylation array analysis for large experiments

Principal Component Analysis for high-dimensional data

Description

Performs principal components analysis on the given gds object and returns the results as an object of class "prcomp".

Usage

## S3 method for class 'gds.class'
prcomp(x, node.name, center = FALSE, scale. = FALSE,
rank. = NULL, retx = FALSE, tol = NULL, perc = 0.01,
npcs = NULL, parallel = NULL, method = c('quick', 'sorted'), verbose = FALSE, ...)

Arguments

x	A gds.class object.
node.name	Name of the gdsn.class node to learn principal components from
center	Logical value indicating whether variables should be shifted to be zero centered.
scale.	Logical value indicating whether the variables should be scaled to have unit variance
tol	a value indicating the magnitude below which components should be omitted.
rank.	(Still functional) Number of principal components to be returned
retx	a logical value indicating whether the rotated variables should be returned.
perc	The percentage of the number of rows that should be used to calculate principle components. Ranging from 0 to 1, a value of 1 would indicicate all rows will be used.
npcs	Number of principal components to be returned
parallel	Can supply either a cluster object (made from makeCluster) or a integer describing the number of cores to be used. This is only used if method="sorted".
method	Indicates whhich method to use out of "quick" and "sorted". "quick" stochastically selects number of rows according to perc. And the supplies them to svd. "sorted" determines the interquartile range for each row then selects the top percentage (according to perc) of probes with the largest interquartile range and supplies selected rows to svd.
verbose	A logical value indicating whether message outputs are displayed.
...	arguments passed to or from other methods. If "x" is a formula one might specify "scale." or "tol".

Details

The calculation is done by a singular value decomposition of the (centered and possibly scaled) data matrix, not by using "eigen" on the covariance matrix. This is generally the preferred method for numerical accuracy. The "print" method for these objects prints the results in a nice format and the "plot" method produces a scree plot.

Value

An object of prcomp class

Examples

data(melon)
e <- es2gds(melon, "melon.gds")
prcomp(e, node.name="betas", perc=0.01, method='quick')
closefn.gds(e)
unlink("melon.gds")

TJGorrie/bigmelon documentation built on Oct. 13, 2023, 9:51 p.m.