Key papers include Albert and Chib (1993)[1] which introduced an approach for
binary and categorical response models based on latent variables that simplifies the Bayesian analysis of categorical response models; Chib and Greenberg (1995)[2] which provided a derivation of the
Metropolis-Hastings algorithm from first principles, guidance on implementation and extensions to multiple-block versions; Chib (1995)[3] where a new method for calculating the marginal likelihood from the Gibbs output is developed; Chib and Jeliazkov (2001)[4] where the method of Chib (1995) is extended to output of Metropolis-Hastings chains; Basu and Chib (2003)[5] for a method for finding marginal likelihoods in Dirichlet process mixture models; Carlin and Chib (1995)[6] which developed a model-space jump method for Bayesian model choice via Markov chain Monte Carlo methods; Chib (1998)[7] which introduced a multiple-change point model that is estimated by the methods of Albert and Chib (1993)[8] and Chib (1996)[9] for hidden Markov processes; Kim, Shephard and Chib (1998)[10] which introduced an efficient inference approach for univariate and multivariate stochastic volatility models;[11][12] and Chib and Greenberg (1998)[13] which developed the Bayesian analysis of the
multivariate probit model.
He has also developed original methods for Bayesian inference in Tobit censored responses,[14] discretely observed diffusions,[15] univariate and multivariate ARMA processes,[16][17] multivariate count responses,[18] causal inference,[19][20] hierarchical models of longitudinal data,[21] nonparametric regression,[22][23] and unconditional and conditional moment models.[24][25]
Key papers include Albert and Chib (1993)[1] which introduced an approach for
binary and categorical response models based on latent variables that simplifies the Bayesian analysis of categorical response models; Chib and Greenberg (1995)[2] which provided a derivation of the
Metropolis-Hastings algorithm from first principles, guidance on implementation and extensions to multiple-block versions; Chib (1995)[3] where a new method for calculating the marginal likelihood from the Gibbs output is developed; Chib and Jeliazkov (2001)[4] where the method of Chib (1995) is extended to output of Metropolis-Hastings chains; Basu and Chib (2003)[5] for a method for finding marginal likelihoods in Dirichlet process mixture models; Carlin and Chib (1995)[6] which developed a model-space jump method for Bayesian model choice via Markov chain Monte Carlo methods; Chib (1998)[7] which introduced a multiple-change point model that is estimated by the methods of Albert and Chib (1993)[8] and Chib (1996)[9] for hidden Markov processes; Kim, Shephard and Chib (1998)[10] which introduced an efficient inference approach for univariate and multivariate stochastic volatility models;[11][12] and Chib and Greenberg (1998)[13] which developed the Bayesian analysis of the
multivariate probit model.
He has also developed original methods for Bayesian inference in Tobit censored responses,[14] discretely observed diffusions,[15] univariate and multivariate ARMA processes,[16][17] multivariate count responses,[18] causal inference,[19][20] hierarchical models of longitudinal data,[21] nonparametric regression,[22][23] and unconditional and conditional moment models.[24][25]