Modeling Zero-Inflated Count Data

When modeling data consisting of counts, it is not uncommon to observe more zeroes than would be expected by typical Poisson or negative binomial distributions. In a zero-inflated count distribution, an additional term is added to account for these extra structural zeroes. This webinar demonstrates 2 new procedures added to Statgraphics 19 for working with zero-inflated distributions, one for fitting such distributions to a sample of data and another for estimating regression models where the dependent variable has extra zeroes.