ChoiceModelR is an open-source package written in the R language by Decision Analyst, Inc. The goal of the ChoiceModelR project is to create a world-class choice modeling software system that:
* Is open-source and free to all
* Includes a broad variety of features and capabilities needed for business, marketing, and scientific applications
* Implements important new techniques continuously
* Runs quickly to handle very large datasets.
We invite interested researchers, students, programmers, and anyone who uses choice modeling, to contribute ideas and R code to the project.
Currently, the ChoiceModelR(tm) package includes the choicemodelr function which implements an MCMC algorithm to estimate a hierarchical multinomial logit model with a normal heterogeneity distribution. The algorithm uses a hybrid Gibbs Sampler with a random walk metropolis step for the MNL coefficients for each unit. Features include:
* Dependent variable may be discrete (nominal or ordinal) or continuous (between zero and one)
* Independent variables may be discrete or continuous
* Constraints may be optionally imposed on estimated coefficients
* Means of the distribution of heterogeneity can optionally be modeled as a linear function of unit characteristics variables
* Number of choice observations per unit (e.g., respondent) may vary within the model data
* Number of choice alternatives per choice observation may vary within the model data.
Possibilities for extending the ChoiceModelR package include:
* New features and capabilities within the choicemodelr function
* New functions for estimating other types of choice models
* New functions for creating experimental designs
* New functions for simulation
You may have other ideas for development. What ideas and needs do you have that you would like to see implemented?
View full history Series and milestones
trunk series is the current focus of development.
