skPartialPCA_step: take a step in sklearn IncrementalPCA partial fit procedure

In vjcitn/BiocSklearn: interface to python sklearn via Rstudio reticulate

take a step in sklearn IncrementalPCA partial fit procedure

Description

take a step in sklearn IncrementalPCA partial fit procedure

Usage

skPartialPCA_step(mat, n_components, obj)

Arguments

mat	a matrix – can be R matrix or numpy.ndarray
n_components	number of PCA to retrieve
obj	sklearn.decomposition.IncrementalPCA instance

Value

trained IncrementalPCA reference, to which 'transform' method can be applied to obtain projection for any compliant input

Note

if obj is missing, the process is initialized with the matrix provided

Examples

# these steps are not basilisk-compliant, you need to acquire references 
irloc = system.file("csv/iris.csv", package="BiocSklearn")
np = reticulate::import("numpy", delay_load=TRUE, convert=FALSE)
irismat = np$genfromtxt(irloc, delimiter=',')
ta = np$take
ipc = skPartialPCA_step(ta(irismat,0:49,0L))
ipc = skPartialPCA_step(ta(irismat,50:99,0L), obj=ipc)
ipc = skPartialPCA_step(ta(irismat,100:149,0L), obj=ipc)
head(names(ipc))
ipc$transform(ta(irismat,0:5,0L))
fullproj = ipc$transform(irismat)
fullpc = prcomp(data.matrix(iris[,1:4]))$x
round(cor(fullpc,fullproj),3)

vjcitn/BiocSklearn documentation built on Feb. 4, 2024, 5:12 a.m.