outlierPlot: Panel of plots displaying multivariate outliers among the...

In abnormally-distributed/cvreg: Cross Validation and Robust Estimation Utilities

Description

Multivariate outliers identify rows of data that contain unusual combinations of values. As such, the definition differs from a univariate outlier as being an extreme value. The value of the ith entry in the jth column may belong to a multivariate outlier, but not be a univariate outlier, and vice versa. This visualizes the distinction. In the displayed plot all variables are centered by their respective medians and scaled by their respective average absolute deviations. This standardization is done only after all calculations are done so that all variables are properly aligned with the y-axis of the plot. The x-axis consists of random coordinates, so the plot may differ if created multiple times. No information is lost or changed. The position along the x-axis is randomized so that the values do not fall on a straight line.

Usage

outlierPlot(x, rmu = NULL, rcov = NULL, plot = T, nrow = 1, ncol = NULL)

Arguments

x	a data frame or matrix of numeric covariates
rmu	robust multivariate center
rcov	robust covariance matrix
plot	if TRUE (the default), the results are plotted using ggplot2. If you desire to make your own plot, plot can be changed to FALSE to return the information used by ggplot2.
nrow	the number of rows for ggplot2's facet_wrap. defaults to 1.
ncol	optional argument to specify the number of columns for ggplot2's facet wrap.

Value

a plot or a data frame.

Examples

out <- cov.ogk(x, all = T)
outlierPlot(x, out$center, out$cov)

abnormally-distributed/cvreg documentation built on May 3, 2020, 3:45 p.m.