Distributions zero inflated Poisson

This section covers introductory theory about the distribution and analysis of count-type data. The focus is on the Poisson and zero-inflated Poisson distributions. The mechanics for implementation in an analysis are also discussed. 

Zero-Inflated Poisson Data Distribution Modei... 

FIGURE 27.4 Poisson and zero-inflated Poisson estimated probability versus observed frequency. The solid circles are the estimated probabilities of observing the episode rate from the Poisson distribution. The open squares are the estimated probabilities of observing the episode rate from the zero-inflated Poisson distribution. The histogram is the distribution of the observed data. 

Table 1 shows the BIG scores for the five probability distributions that were fitted to all attacks, and to those countries that have a share of at least 2.5% of all attacks. The Negative Binominal distribution (NBI) fits clearly better than the Poisson distribution, as indicated by the substantially lower BIG scores. The Geometric distribution (GEO), which is a special case of the NBI, performs practically the same. Similarly, the Zero-Inflated (ZIP) and Zero-Adjusted (ZAP) Poisson models are only marginally different from the POL The NBI has the following advantages compared to the POL... 

---