Empirical Examination of the Poisson Regression Residuals for the Evaluation of Influential Points

Khan, Aamna; Ullah, Muhammad Aman; Amin, Muhammad; Muse, Abdisalam Hassan; Aldallal, Ramy; Mohamed, Mohamed S.

doi:10.1155/2022/6995911

Cited by 1 publication

(1 citation statement)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The Poisson regression model assumes that the counts follow a Poisson distribution. It is suitable when the mean and variance of the counts are approximately equal [19,20]. The Poisson model can be expressed as:…”

Section: Poisson Regression Modelmentioning

confidence: 99%

Unraveling New Insights: Exploring Risk Factors of Under-Five Mortality in Somaliland using the First-Ever SLDHS 2020 Data and Count Data Regression Models

Ismail,

Chesneau,

Muse

2023

Preprint

Self Cite

View full text Add to dashboard Cite

The high prevalence of zero counts in child mortality distribution in Somaliland prompts an exploration into whether distinct generating processes exist for child mortality and their frequency per mother. This study aims to identify an appropriate count regression model for the number of children who died per mother in Somaliland, addressing concerns of under-dispersion and zero-inflation. Addition[1]ally, it seeks to determine the risk factors associated with under-five mortality and their frequency. The analysis utilizes data from the nationally representative 2020 Somaliland Demographic and Health Survey (SLDHS), where 28.1% of the 4,615 surveyed women reported no under-five mortality. These zero counts may arise from structural or sampling factors, leading to the investigation of two-part zero-inflated and hurdle-count regression models alongside standard one-part count regression models. Diagnostic tools, such as the likelihood ratio test, The Akaike information criterion (AIC) and the Bayesian information criterion (BIC), are employed to determine the most suitable model. The hurdle Poisson regression model emerges as the best fit. The findings highlight several significant factors associated with under-five mortality per mother, including child sex, number of children ever born, access to clean water sources and toilet facilities, geographic region, wealth index, breastfeeding, mothers’ age at first birth, and mothers’ education. Understanding and addressing these factors can guide targeted interventions and policies aimed at reducing under-five mortality and improving child health outcomes. Furthermore, promoting deliveries at healthcare institutions is recommended for mothers.

show abstract

Section: Poisson Regression Modelmentioning

confidence: 99%