outliers: Outliers detection on a variable using the adjusted...

In ldbk/fishPifct: fishPi functions

Description

Outliers detection on a variable using the adjusted outlyingness index (based on the robustbase package) if the variable is numeric, and on a frequency table if not. In the later case, a treshold in percentage is used to flag an outlier (1 treshold, then it is reported as outlier).

Usage

outliers(obj, slot, var, treshold = 1, ...)

Arguments

obj:	a csPi object
slot:	the slot where the variable of interest is located
var:	the name of a variable on which outiers detection is applied
treshold:	the percentage treshold to consider outliers in the case of non numeric data
other	parameters related to the function adjOutlyingness from the package robustbase

Value

plot a graphics and return a table of the table individuals flagged as outliers

Author(s)

Laurent Dubroca

Examples

## Not run: 
 data(sole)
 sole<-csDataTocsPi(sole.cs)
 outliers(sole,"foDur")

## End(Not run)

ldbk/fishPifct documentation built on May 20, 2019, 11:28 p.m.