A simulation study for count data models under varying degrees of outliers and zeros

Abstract

This study was aimed at examining the performance of count data models under various outliers and zero inflation situations with simulated data. Poisson, Negative Binomial, Zero-inflated Poisson, Zero-inflated Negative Binomial, Poisson Hurdle and Negative Binomial Hurdle models were considered to test how well each of the model fits the selected datasets having outliers and excess zeros. We found that Zero-inflated Negative Binomial and Negative Binomial Hurdle models were found to be more successful than other count data models. Also the results indicated that in some scenarios, the Negative Binomial model outperformed other models in the presence of outliers and/or excess zeros.

Keywords:

• Count data
• Hurdle models
• Outliers
• Zero-inflated models