isOutlier: Find multivariate outliers (Leys et al., 2018, JESP)

In akcochrane/ACmisc: Aaron Cochrane's miscellaneous functions

Find multivariate outliers (Leys et al., 2018, JESP)

Description

Tests for multivariate outliers using a robust method (cited above). In summary, a minimum proportion of cases are used to estimate the location and scatter of the multivariate space, then that location and scatter are used to test for [multivariate] outlying values.

Usage

isOutlier(dat, dat_proportion = "check", reject_alpha = 0.01)

Arguments

dat	data
dat_proportion	Proportion of data to retain in estimating the robust multivariate location and scatter. Defaults to testing at various proportions.
reject_alpha	P-value threshold for excluding cases (using a chi-square test on Mahalanobis distances; Leys et al, 2018)

Details

Takes a numeric data frame. Returns a logical vector indicating whether the cases should be excluded.

Is NOT robust to missing data.

See Also

cov.rob for estimating the robust location and scatter.

akcochrane/ACmisc documentation built on Nov. 24, 2024, 11:22 a.m.