Tools for influence diagnostics in statistical models

Set of routines for influence diagnostics by using case-deletion in ordinary least squares, ridge estimation and LAD regression.

Features

Initial release (Oct 29, 2023) of india package have implemented the following influence measures for objects of class lad, ols and ridge, provided by L1pack and fastmatrix, respectively.

• Cook's distances.
• Leverages (and hatvalues).
• Likelihood displacement.
• Relative change in the condition number.

Our plan in the near future is the implementation of functions to handle:

• General models commonly used in Statistics.
• Procedures for assessing the local influence considering several perturbation schemes.

Reference Manual

• india.pdf

Resources

Binaries and sources for india are available here (these are local files):

• india_0.1.tar.gz - Package sources
• india_0.1.zip - Windows binaries

Version 0.1 of india can be found at the CRAN package repository:

• india_0.1.tar.gz - Package sources
• india_0.1.tgz - MacOS binaries (R-release, x86_64)

Installation instructions

To install india (version 0.1) from CRAN, start R and enter:

install.packages("india")

Or install it from its GitHub repository. First install the devtools package.

install.packages("devtools")

Then install india using the install_github function in devtools

library(devtools)
install_github("faosorios/india", subdir = "pkg")

Alternatively, you can download the source as a tarball (.tar.gz file). Unpack this file (thereby creating a directory named, india) and install the package source by executing (at the console prompt)

R CMD INSTALL india

Next, you can load the package by using the command: library(india)

Providing Feedback

Please report any bugs/suggestions/improvements to Felipe Osorio. If you find these routines useful or not then please let me know. Also, acknowledgement of the use of the routines is appreciated.

To cite the fastmatrix package in publications use:

Osorio, F. (2023). india: Influence diagnostics in statistical models. R package version 0.1. URL: https://cran.r-project.org/package=india

About the Author

Felipe Osorio is an Assistant Professor at Department of Mathematics, Universidad Tecnica Federico Santa Maria, Chile.

• Webpage: fosorios.mat.utfsm.cl