Modelling Zero-Inflated Over Dispersed Dengue Data Via Zero-Inflated Poisson Inverse Gaussian Regression Model: A Case Study Of Bangladesh

Bangladesh has been noted for experiencing some of the most susceptible dengue outbreaks in Asia; the country’s location, dense population, and changing environment all play major roles in this. Determining the correlation between meteorological conditions and case count is critical for predicting about the characteristics of the DENV outbreak. Certain widely used models, such as the Poisson regression model or the negative binomial regression model, are insufficient to adequately predict dengue fever since many of these datasets are of the over-dispersed, long-tail variety, and zero-inflated. In this study, the Zero-inflated negative binomial regression model is compared with the Zero-inflated Poisson inverse Gaussian regression model. Depending on AIC and BIC criteria Zero-inflated Poisson inverse-Gaussian regression model is proposed. Then Zero-inflated Poisson inverse Gaussian regression model is used to model the dataset containing confirmed positive cases of dengue fever and seven meteorological variables. The proposed model shows that all the meteorological variables are significantly associated with the confirmed positive cases of dengue fever. That’s why modeling a dengue-fever dataset with a Zero-inflated Poisson-inverse Gaussian regression model is suggested in this study.