ds.PCA: Principal component analysis (PCA)

In isglobal-brge/dsMLClient: DataSHIELD client site machine learning functions

Principal component analysis (PCA)

Description

Perform a non-disclosive PCA with the option of centering and scaling (unit variance) the pooled data

Usage

ds.PCA(
  x,
  method = "fast",
  ncomp = 2,
  scale = TRUE,
  center = TRUE,
  datasources = NULL
)

Arguments

x	character Name of the data frame on the server that has the dataset
method	character (default "fast") Method of the SVD calculation. "fast" to do a truncated SVD of the first ncomp components, "full" to do the complete SVD calculation.
ncomp	numeric (default 2) Number of components to calculate when method = "fast"
scale	bool (default TRUE) Set to TRUE to scale the dataset (pooled approach), FALSE to not scale
center	bool (default TRUE) Set to TRUE to center the dataset (pooled approach), FALSE to not scale
datasources	a list of DSConnection-class (default NULL) objects obtained after login

Value

data frame With the principal components (columns) and variables (rows)

isglobal-brge/dsMLClient documentation built on March 14, 2023, 1:59 p.m.