Fast Estimation of Multinomial Logit Models: R Package mnlogit

Asad Hasan, Zhiyu Wang and Alireza S. Mahani

Journal of Statistical Software, 2016, vol. 075, issue i03

Abstract: We present the R package mnlogit for estimating multinomial logistic regression models, particularly those involving a large number of categories and variables. Compared to existing software, mnlogit offers speedups of 10 - 50 times for modestly sized problems and more than 100 times for larger problems. Running in parallel mode on a multicore machine gives up to 4 times additional speedup on 8 processor cores. mnlogit achieves its computational efficiency by drastically speeding up computation of the log-likelihood function's Hessian matrix through exploiting structure in matrices that arise in intermediate calculations. This efficient exploitation of intermediate data structures allows mnlogit to utilize system memory much more efficiently, such that for most applications mnlogit requires less memory than comparable software by a factor that is proportional to the number of model categories.

Date: 2016-11-19
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (4)

Downloads: (external link)
https://www.jstatsoft.org/index.php/jss/article/view/v075i03/v75i03.pdf
https://www.jstatsoft.org/index.php/jss/article/do ... mnlogit_1.2.5.tar.gz
https://www.jstatsoft.org/index.php/jss/article/do ... ile/v075i03/v75i03.R
https://www.jstatsoft.org/index.php/jss/article/do ... i03/simChoiceModel.R

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:jss:jstsof:v:075:i03

DOI: 10.18637/jss.v075.i03

Access Statistics for this article

Journal of Statistical Software is currently edited by Bettina Grün, Edzer Pebesma and Achim Zeileis

More articles in Journal of Statistical Software from Foundation for Open Access Statistics
Bibliographic data for series maintained by Christopher F. Baum ().