Zero-Inflated Negative Binomial Regression

The zero-inflated negative binomial (ZINB) model in PROC TCOUNTREG is based on the negative binomial model with quadratic variance function (p=2). The ZINB model is obtained by specifying a negative binomial distribution for the data generation process referred to earlier as Process 2:

     

Thus the ZINB model is defined to be

     
     
     

In this case, the conditional expectation and conditional variance of are

     
     

As with the ZIP model, the ZINB model exhibits overdispersion because the conditional variance exceeds the conditional mean.

ZINB Model with Logistic Link Function

In this model, the probability is given by the logistic function—namely,

     

The log-likelihood function is

     
     
     
     

See the section Poisson Regression for the definition of .

The gradient for this model is given by

     
     
     
     
     
     

ZINB Model with Standard Normal Link Function

For this model, the probability is specified with the standard normal distribution function (probit function): . The log-likelihood function is

     
     
     
     
     
     
     

See the section Poisson Regression for the definition of .

The gradient for this model is given by

     
     
     
     
     
     

Note: This procedure is experimental.