Constructs non-parametric bootstrap Bias-corrected and accelerated (BCa) confidence limits in Principal Component Analysis for scores and loadings. Further outputs include bootstrap estimates, standard error, and bias. Bootstrap confidence limits are useful to evaluate uncertainty associated with samples or variables for outlier detection and variable selection.
Hamid Babamoradi [1]Frans W.J. van den Berg [1]Asmund Rinnan [1]
[1]Department of Food Science, Spectroscopy & Chemometrics Section, University of Copenhagen, Denmark
