Computes statisticsl limits for orthogonal and score distances (based on calibration set) and assign the calculated values as model properties.
# S3 method for pca
setDistanceLimits(
obj,
lim.type = obj$lim.type,
alpha = obj$alpha,
gamma = obj$gamma,
...
)
Object models with the two fields updated.
object with PCA model
type of limits ("jm", "chisq", "ddmoments", "ddrobust")
significance level for detection of extreme objects
significance level for detection of outliers (for data driven approach)
other arguments
The limits can be accessed as fields of model objects: $Qlim
and $T2lim
. Each
is a matrix with four rows and ncomp
columns. First row contains critical limits for
extremes, second row - for outliers, third row contains mean value for corresponding distance
(or its robust estimate in case of lim.type = "ddrobust"
) and last row contains the
degrees of freedom.