WALS

CRANWALSMaRDI QIDQ5980388

Weighted-Average Least Squares Model Averaging

Kevin Huynh

Last update: 3 November 2023

Copyright license: GNU General Public License, version 3.0, GNU General Public License, version 2.0

Software version identifier: 0.2.3, 0.2.4

Description

Implements Weighted-Average Least Squares model averaging for negative binomial regression models of Huynh (2023) (mimeo), generalized linear models of De Luca, Magnus, Peracchi (2018) <doi:10.1016/j.jeconom.2017.12.007> and linear regression models of Magnus, Powell, Pruefer (2010) <doi:10.1016/j.jeconom.2009.07.004>, see also Magnus, De Luca (2016) <doi:10.1111/joes.12094>. Weighted-Average Least Squares for the linear regression model is based on the original 'MATLAB' code by Magnus and De Luca <https://www.janmagnus.nl/items/WALS.pdf>, see also Kumar, Magnus (2013) <doi:10.1007/s13571-013-0060-9> and De Luca, Magnus (2011) <doi:10.1177/1536867X1201100402>.

Cites work

• Weighted-average least squares estimation of generalized linear models
• A characterization of Bayesian robustness for a normal location parameter
• A comparison of two model averaging techniques with an application to growth empirics
• WEIGHTED‐AVERAGE LEAST SQUARES (WALS): A SURVEY
• Bayesian Model Averaging and Weighted-Average Least Squares: Equivariance, Stability, and Numerical Issues

This page was built for software: WALS