ipca_mono: interface to incremental PCA from sklearn: mono implies...

In vjcitn/useret: demo of reticulate interface in a package

Description

interface to incremental PCA from sklearn: mono implies monolithic data handling

Usage

ipca_mono(x, n_comp = 2L, batch_size = 25L, ...)

Arguments

x	either character string to HDF5 matrix or R matrix
n_comp	integer number of PC to compute
batch_size
...	not used

Examples

d = data.matrix(iris[,1:4])
p = prcomp(d)
dim(p$x)
h5targ = tempfile()
require(rhdf5)
h5createFile(h5targ)
h5write(d, h5targ, "iris")
tst = ipca_mono(h5targ, n_comp=4L)
tstm = tst@fit_transform
colnames(tstm) = paste0("ip", 1:4)
pairs(cbind(p$x, tstm))

vjcitn/useret documentation built on May 8, 2019, 11:19 p.m.