Power Analysis for PLS Classification.
powerPLS
This package provides a tool to perform power analysis in Partial Least Squares (PLS) for classification when two classes are analyzed.
Installation
You can install the released version of powerPLS
with:
devtools::install_github("angeella/powerPLS")
Quick overview
The main functions are
computeSampleSize()
which estimated the power considering several values of sample size and number of score components.computePower()
which estimated the power considering a fixed sample size and several number of score components.
datas <- simulatePilotData(nvar = 30, clus.size = c(5,5),m = 6,nvar_rel = 5,ncomp = 2)
out <- computePower(X = datas$X, Y = datas$Y, A = 3)
out <- computeSampleSize(X = datas$X, Y = datas$Y, A = 3, n = c(10,20,30))
References
Andreella, A., Finos, L., Scarpa, B. and Stocchero, M. "Towards a power analysis for PLS-based methods" arXiv:2403.10289 stat.ME. link: https://arxiv.org/abs/2403.10289
Did you find some bugs?
Please write to angela.andreella[\at]unive[\dot]it or insert a reproducible example using reprex on my issue github page.