Zero inflated models (frequentist and Bayesian)

Introduction to Zero Inflated Models   

-Bayesian and frequentist approaches-


The course material is accessible via the menu on the left.

Suppose you want to study hippos and the effect of habitat variables on their distribution. When sampling, you may count zero hippos at many sites, potentially resulting in overdispersed Poisson GLMs.  In such cases zero inflated models can be applied. During the course several case studies are presented, in which the statistical theory for zero inflated models is integrated with applied analyses in a clear and understandable manner. Zero inflated models consist of two integrated GLMs and therefore we will start with a revision of GLM.

Zero inflated GLMMs for nested data (repeated measurements, short time series, clustered data, etc.) are discussed in the second part of the course. We will focus on zero inflated count data, and zero inflated continuous data.


Zero inflated count data. Zero inflated continuous data. Dependency. ZIP and ZAP models. Zero inflated GLMMs with random effects. Bayesian statistics, MCMC and JAGS. lme4, glmmADMB, JAGS. Overdispersion and solutions. Bayesian model selection.