Estimation Tools Review - ActivitySim/activitysim GitHub Wiki

The purpose of this page is to review available Python estimation tools in order to select one for prototyping estimation integration. The activitysim sub-models to be integrated and estimated at this point in time are auto ownership, work location, and tour mode choice. Auto ownership is a multinomial logit model with few alternatives, work location is an multinomial logit model with size terms and many alternatives, and tour mode choice is a nested logit model with few alternatives but a significant amount of data.

Feature	larch	PandasBiogeme	pylogit	choicemodels
Includes multinomial logit	x	x	x	x
Includes nested logit	x	x	x
Efficient handling of large alternative sets	x	x		x
Active development	x			x
Sufficient user community	x	x	x	x
Good documentation	x	x	x	x
Industry maintainer	x	x
Focus on practical models	x		x	x

We decided to use larch for prototyping integration because it supports nested logit, supports large alternative sets for destination choice, is being actively maintained and developed, has good documentation, is maintained by a member of the industry, and is oriented to practice.