Discrete Choice (Binary, Poisson and Ordered) Models with Random Parameters

An implementation of simulated maximum likelihood method for the estimation of Binary (Probit and Logit), Ordered (Probit and Logit) and Poisson models with random parameters for cross-sectional and longitudinal data as presented in Sarrias (2016) .

Rchoice

Rchoice is a package in R for estimating Ordered, Binary and Poisson models with random parameters for cross-sectional and panel data.

What kind of models can be estimated?

• Binary (Logit/Probit), Ordered (Logit/Probit) and Poisson Models with fixed (non-stochastic) parameters.
• Binary (Logit/Probit), Ordered (Logit/Probit) and Poisson models with random coefficients. The distribution of the coefficients can be normal, log-normal, truncated normal, triangular, uniform and Johnson Sb.
• Random Effects model for Panel or longitudinal data.
• Estimate the conditional individual-specific coefficient.
• Binary models with (Logit/Probit) with heteroskedasticity using MLE and compute the marginal effects.
• Probit model with endogenous continuous covariate using MLE and compute marginal effects.