In rgcstats/surveyoutliers: Helps Manage Outliers in Sample Surveys

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "README-"
)

surveyoutliers

This package contains various procedures to deal with outliers in survey data. The only method available at present is to calculate optimal winsorising cutoffs as described in Kokic and Bell (1994) and Clark (1995).

Installation

library(devtools)
devtools::install_github("rgcstats/surveyoutliers")

Quick demo

library(surveyoutliers)
# The dataset test contains a variable of interest y, and generalized regression
#  weights regwt calculated using auxiliary variables x1 and x2,
#  and inverse probabilities of selection piwt
# The following code calculates the optimal tuning parameter and hence winsorised values.
test <- optimal.onesided.cutoff(formula=y~x1+x2,surveydata=survdat.example)
# Show the results by plotting the winsorised vs the raw values of the variable of interest:
plot(test$windata$y,test$windata$win1.values)

References

Clark, R. G. (1995), "Winsorisation methods in sample surveys," Masters thesis, Australian National University, http://hdl.handle.net/10440/1031.

Kokic, P. and Bell, P. (1994), "Optimal winsorizing cutoffs for a stratified finite population estimator," J. Off. Stat., 10, 419-435.

rgcstats/surveyoutliers documentation built on May 27, 2019, 7:22 a.m.