ESD_test {mcradds}

EDS Test for Outliers

Description

Perform Rosner's generalized extreme Studentized deviate (ESD) test, which assumes that the distribution is normal (Gaussian), can be used when the number of outliers is unknown, and becomes more robust as the number of samples increases.

Usage

ESD_test(x, alpha = 0.05, h = 5)

Arguments

x

(numeric)
vector of observations that can be the difference from Bland-Altman analysis. Normally the relative difference is preferred in IVD trials. Missing(NA) is allowed but will be removed. There must be at least 10 available observations in x.

alpha

(numeric)
type-I-risk, \alpha.

h

(integer)
the positive integer indicating the number of suspected outliers. The argument h must be between 1 and n-2 where n denotes the number of available values in x. The default value is h = 5.

Value

A list class containing the results of the ESD test.

• stat a data frame contains the several statistics about ESD test that includes the index(i), Mean, SD, raw data(x), the location(Obs) in x, ESD statistics(ESDi), Lambda and Outliers(TRUE or FALSE).

• ord a vector with the order index of outliers that is equal to Obs in the stat data frame.

Note

The algorithm for determining the number of outliers is as follows:

• Compare ESDi with Lambda. If ESDi > Lambda then the observations will be regards as outliers.

• The order index corresponds to the available x data that has been removed the missing (NA) value.

• As we should compare if the ESD(h) and ESD(h+1) are equal, the h+1 ESD values will be shown. If they are identical, both of them can not be regarded as outliers.

References

CLSI EP09A3 Appendix B. Detecting Aberrant Results (Outliers).

Examples

data("platelet")
res <- blandAltman(x = platelet$Comparative, y = platelet$Candidate)
ESD_test(x = res@stat$relative_diff)