Calculation of Fisher Information Matrix and/or Asymptotic Variance-Covariance Matrix

[Rachel Murray-Watson]

Hi all, 

I have just run a pilot study and want to ensure that the sample size for my main study is sufficient. According to many of the methods listed in this paper, I need to have a asymptotic variance-covariance matrix (or the Fisher's Information Matrix to calculate the AVC). Can anyone advise as to how to do this using the outputs from biogeme? 

Paper: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4575371/#CR21


Best wishes, 

Rachel 

[Michel Bierlaire]

Have a look at Appendix B of this document: https://transp-or.epfl.ch/documents/technicalReports/Bier23.pdf

> --
> You received this message because you are subscribed to the Google Groups "Biogeme" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to biogeme+u...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/biogeme/ee9dde78-0dd3-40df-8154-8d7bcd6cbaa5n%40googlegroups.com.

Michel Bierlaire
Transport and Mobility Laboratory
School of Architecture, Civil and Environmental Engineering
EPFL - Ecole Polytechnique Fédérale de Lausanne
http://transp-or.epfl.ch
http://people.epfl.ch/michel.bierlaire

[Rachel Murray-Watson]

Ah great, thank you! So when it says "Biogeme reports statistics based on both the Cramer-Rao estimate (18) and the robust estimate", does that mean that results.getVarCovar() returns the AVC? 

[Michel Bierlaire]

results.getVarCovar --> Rao-Cramer
results.getRobustVarCovar --> the robust version

> To view this discussion on the web visit https://groups.google.com/d/msgid/biogeme/1eebc90a-a93a-4f10-a895-6650fd642c9en%40googlegroups.com.