This PR moves all the ChoiceModels work from the Urban Analytics Lab (UAL) fork up to the main repository. It includes the following:
Code pulled from urbansim.urbanchoice (with complete commit history)
MergedChoiceTable() class for generating a merged long-format table of choosers and alternatives -- this is a new, generalized interface for urbansim.urbanchoice.interaction.mnl_interaction_dataset()
MultinomialLogit() class for estimating MNL models, using either the PyLogit engine or the engine from urbansim.urbanchoice.mnl.mnl_estimate()
MultinomialLogitResults() class for storing and reporting fitted MNL models, including ability to print a Statsmodels-style results table for models estimated with either engine
MNL demo data derived from the public California Household Travel Survey tables
This PR moves all the ChoiceModels work from the Urban Analytics Lab (UAL) fork up to the main repository. It includes the following:
Code pulled from
urbansim.urbanchoice(with complete commit history)MergedChoiceTable()class for generating a merged long-format table of choosers and alternatives -- this is a new, generalized interface forurbansim.urbanchoice.interaction.mnl_interaction_dataset()MultinomialLogit()class for estimating MNL models, using either the PyLogit engine or the engine fromurbansim.urbanchoice.mnl.mnl_estimate()MultinomialLogitResults()class for storing and reporting fitted MNL models, including ability to print a Statsmodels-style results table for models estimated with either engineMNL demo data derived from the public California Household Travel Survey tables
Various work and demos in Jupyter notebooks