buildPCA: buildPCA

In SPOT: Sequential Parameter Optimization Toolbox

buildPCA

Description

buildPCA builds principal components of given dataset. It is used inside plotPCA function to build necessary object to perform principal components analysis.

Usage

buildPCA(x, control = list())

Arguments

x	dataset of parameters to be transformed
control	control list

Value

returns a list with the following elements:
sdev the standard deviations of the principal components (i.e., the square roots of the eigenvalues of the covariance/correlation matrix, though the calculation is actually done with the singular values of the data matrix).
rotation the matrix of variable loadings (i.e., a matrix whose columns contain the eigenvectors).
x transformed matrix.
center,scale the centering and scaling used, or FALSE.

Author(s)

Alpar Gür [e-mail]

Examples

#define objective function

objFun <- function(x) 2*(x[1] - 1)^2 + 5*(x[2] - 3)^2 + (10*x[3] - x[4]/3)   

spotConfig <- 
list(types = c('numeric', 'numeric', 'numeric', 'numeric'), 
funEvals = 15, #budget
noise = TRUE,
seedFun = 1,
replicated = 2,
seedSPOT = 1,
design = designLHD,
model = buildRandomForest, #surrogate model
optimizer = optimLHD, #LHD to optimize model
optimizerControl = list(funEvals=100)) #100 model evals in each step

lower <- c(-20, -20, -20, -20) 
upper <- c(20, 20, 20, 20)

res <- spot(x=NULL, 
   fun=objFun, 
   lower=lower, 
   upper=upper, 
   control=spotConfig) 

resPCA <- buildPCA(res$x) 

SPOT documentation built on June 26, 2022, 1:06 a.m.