DCM: Discrete choice models
Status: work in progress.
This page is related to Discrete Choice Models (DCM) based on random utility maximization approach (RUM). This project was accepted for GSoC 2013.
Proposal: http://www.google-melange.com/gsoc/proposal/review/google/gsoc2013/anamp/1
Blog Updates: http://gsocstatsmodels.blogspot.com.es/
We are working on Multinomial Logit Model which variables could vary over alternatives (also called Conditional Logit Model). See: https://github.com/AnaMP/statsmodels/compare/clogit
You can see an example of use here: http://nbviewer.ipython.org/6564526
If you try it, please, let me know any comment.
Is planned to work on the nested logit and mixed logit algorithms.
You can see an outline with:
- cases of use, properties and references of the principal DCM based on RUM.
- statistics software packages and source codes for DCM estimation.
here: https://docs.google.com/spreadsheet/pub?key=0AsJlEo80UF54dDdTNXFCYUpzdDZ1eVBxOWU0OTgzMkE&gid=3
General References ----------Greene, W. Econometric Analysis. Prentice Hall, 5th. edition. 2003
Train, K. Discrete Choice Methods with Simulation. Cambridge University Press. 2003