Generalized Poisson Regression Research Articles

Analyzing Number of Children Ever Born in Bangladesh Using Generalized Poisson Regression Model

One major indicator to detect any country’s future population size is the total fertility rate (TFR). Based on the total number of children ever born, an underdispersed count data, this paper aims to investigate the fertility behavior of mothers in Bangladesh who have given birth to one or more children. For analyzing count dataset, the Poisson regression model (PRM) is commonly used, which assumes that the response variable’s mean is equal to its variance. Since the generalized Poisson regression model (GPRM) is suitable for the analysis of both overdispersed and underdispersed count data, this model is applied to deal with the underlying count dataset in this study. The results obtained in this paper have revealed some variables that impose significant impact on the current fertility differential of mothers giving birth to children. Dhaka Univ. J. Sci. 72(2): 01-06, 2024 (July)

Achieving food security: Household perception and adoption of home gardening techniques in Ghana

As global food crises and high food prices continue to impact communities, there is a growing focus on strengthening local food systems and improving livelihoods. Home gardening techniques have gained renewed attention as a means of food production and enhancing household food security. However, there is a lack of empirical studies on the perception and adoption of home gardening techniques. This study therefore fills this gap by examining the perception and adoption of home gardening techniques among households in Ghana. A multistage sampling technique was used to select 348 households and data was collected through a structured questionnaire. Using a 5-point Likert scale, we analyzed the households' perception of home gardening techniques, and utilized a multivariate probit regression model to examine the factors influencing multiple adoption of home gardening techniques. The Kendall’s coefficient of concordance was used to test the level of agreement among ranked constraints in adopting home gardening techniques. The generalized Poisson regression was used to examine the number of adoption of home gardening techniques. The results revealed that most households perceived home gardening techniques as a way to improve household food security. Households were found to adopt multiple home gardening techniques, such as use of pots, sacks, and tires. Empirically, factors such as sex, education, occupation, income, technology cost, and social network had a significant influence on multiple adoption of home gardening techniques. Also, age, education, household size, income, technology cost, social network, inadequate space and insect and disease damages significantly influenced the number of adoptions of home gardening techniques. Furthermore, lack of time and inadequate space were the main constraints limiting the adoption of these techniques. We suggest government interventions to spur the adoption of home gardening techniques to improve household food nutrition.

Simulation Analysis of Asymptotic Normality of Maximum Likelihood Estimation Based on the Fisher Scoring Algorithm in Generalized Poisson Regression Modeling

Simulation Analysis of Asymptotic Normality of Maximum Likelihood Estimation Based on the Fisher Scoring Algorithm in Generalized Poisson Regression Modeling

Cultivating sustainability: Adoption and intensity of soil fertility management technologies among rural farms in Zambia

Despite the numerous advantages of soil fertility management (SFM) practices and substantial investments to promote adoption, uptake of these technologies remains low across sub-Saharan Africa. We study the adoption and intensity of key SFM practices - improved fallow, animal manure, compost, crop residue retention, minimum tillage, inorganic fertilizer, and intercropping among rural farmers in Zambia. Data from 1234 rural farms and the multivariate probit (MVP) and generalized Poisson regression models are used. The MVP model highlights interconnections among practices, emphasizing the need for integrated interventions. Socio-demographics like gender, age, education, and household labor availability significantly influence adoption of SFM practices. Farm characteristics including experience, land ownership, livestock, and off-farm income also play pivotal roles. Accessibility of information via phones, radio, and farmer groups emerged as a crucial enabler, while extension services, credit, and proximity to markets and roads shape adoption decisions. Results indicate education, household labor, off-farm income, and group membership drive intensity of adoption. Extension contacts, training, and agro-ecological region location also affect intensity with regional variations. Key policy implications emerge. First, the complementarities among practices underscore the need for holistic strategies recognizing these synergies. Second, enhancing education specifically for younger farmers can promote adoption of SFM technologies by increasing knowledge. Third, spurring farmer groups facilitates knowledge exchange, input access, and collaborative adoption. Fourth, strategic extension services and training programs are vital to address knowledge gaps for sustained adoption. Overall, this study provides insights to guide policies for promoting SFM practices among rural farms.

Factors associated with exclusive breastfeeding by maternal HIV status: a population-based survey in Kenya

BackgroundExclusive breastfeeding (EBF) in the first six months remains low globally, despite known benefits of lower morbidity and mortality among breastfed infants. It is important to understand factors associated with breastfeeding to support optimal breastfeeding practices, particularly in settings with a high burden of HIV.MethodsWe analyzed data from a population-level survey of mother-infant pairs attending 6-week or 9-month immunizations at 141 clinics across Kenya. Primary outcomes included maternal report of (1) EBF at 6-week visit, defined as currently feeding the infant breast milk only, (2) EBF for the first 6-months of life, defined as breastfeeding or feeding the infant breast milk only with no introduction of other liquids or solid foods until 6 months, and (3) continued breastfeeding with complementary feeding at 9-months. Correlates of breastfeeding practices were assessed using generalized Poisson regression models accounting for facility-level clustering.ResultsAmong 1662 mothers at 6-weeks, nearly all self-reported breastfeeding of whom 93% were EBF. Among 1180 mothers at 9-months, 99% had ever breastfed, 94% were currently breastfeeding and 73% reported 6-month EBF. At 6-weeks, younger age (< 25 years) (adjusted Prevalence Ratio (aPR) 0.96; 95% CI 0.93, 0.99), lower education (aPR 0.96; 95% CI 0.93, 0.99) and recent infant illness (aPR 0.97; 95% CI 0.94, 1.00) were associated with lower EBF prevalence while women living with HIV (WLWH) had higher EBF prevalence (aPR 1.06; 95% CI 1.02, 1.10) than women without HIV. 6-month EBF prevalence was 26% higher in WLWH (aPR 1.26; 95% CI 1.15, 1.35) than women without HIV, 14% lower in women reporting mild or above depressive symptoms (aPR 0.86; 95% CI 0.76, 0.99) than those with none or minimal depressive symptoms, and 15% lower in women with versus without history of intimate partner violence (aPR 0.85; 95% CI 0.74, 0.98). At 9-months, WLWH had a lower prevalence of continued breastfeeding with complementary feeding (aPR 0.73; 95% CI 0.64, 0.84) than women without HIV.ConclusionWLWH had higher EBF prevalence in the first 6-months, but lower prevalence of continued breastfeeding at 9-months. Strategies to support EBF and continued breastfeeding beyond 6-months postpartum, particularly among WLWH, are needed.

The impact of apparent temperature on the emergency visits for traumatic fractures in Hangzhou, China

BackgroundTraumatic fractures occur frequently worldwide. However, research remains limited on the association between short-term exposure to temperature and traumatic fractures. This study aims to explore the impact of apparent temperature (AT) on emergency visits (EVs) due to traumatic fractures.MethodsBased on EVs data for traumatic fractures and the contemporary meteorological data, a generalized Poisson regression model along with a distributed lag nonlinear model (DLNM) were undertaken to determine the impact of AT on traumatic fracture EVs. Subgroup analysis by gender and age and sensitivity analysis were also performed.ResultsA total of 25,094 EVs for traumatic fractures were included in the study. We observed a wide “J”-shaped relationship between AT and risk of traumatic fractures, with AT above 9.5 °C positively associated with EVs due to traumatic fractures. The heat effects became significant at cumulative lag 0–11 days, and the relative risk (RR) for moderate heat (95th percentile, 35.7 °C) and extreme heat (99.5th percentile, 38.8 °C) effect was 1.311 (95% CI: 1.132–1.518) and 1.418 (95% CI: 1.191–1.688) at cumulative lag 0–14 days, respectively. The cold effects were consistently non-significant on single or cumulative lag days across 0–14 days. The heat effects were higher among male and those aged 18–65 years old. The sensitivity analysis results remained robust.ConclusionHigher AT is associated with cumulative and delayed higher traumatic fracture EVs. The male and those aged 18–65 years are more susceptible to higher AT.

Mapping QTL controlling count traits with excess zeros and ones using a zero-and-one-inflated generalized Poisson regression model.

The research on the quantitative trait locus (QTL) mapping of count data has aroused the wide attention of researchers. There are frequent problems in applied research that limit the application of the conventional Poisson model in the analysis of count phenotypes, which include the overdispersion and excess zeros and ones. In this article, a novel model, that is, the zero-and-one-inflated generalized Poisson (ZOIGP) model, is proposed to deal with these problems. Based on the proposed model, a score test is performed for the inflation parameter, in which the ZOIGP model with a constant proportion of excess zeros and ones is compared with a standard generalized Poisson model. To illustrate the practicability of the ZOIGP model, we extend it to the QTL interval mapping application that underpins count phenotype with excess zeros and excess ones. The genetic effects are estimated utilizing the expectation-maximization algorithm embedded with the Newton-Raphson algorithm, and the genome-wide scan and likelihood ratio test is performed to map and test the potential QTLs. The statistical properties exhibited by the proposed method are investigated through simulation. Finally, a real data analysis example is used to illustrate the utility of the proposed method for QTLmapping.

Comparison of Geographically Weighted Artificial Neural Network and Geographically Weighted Generalized Poisson Regression on Crime Cases in East Java Indonesia

Crime is behavior that violates laws and social norms. One area of Indonesia with a relatively high number of crimes is East Java. Data on the number of crimes in East Java in 2020 showed overdispersion and multicollinearity. This study aims to model the number of crimes reported in East Java by considering the spatial effects in the data. The methods used to analyze the data are GPR, GWANN and GWGPR. Moreover, we also determine the results of comparing the three methods using R2 and RMSE. The study results show that GWANN provides better results to model the number of reported crimes compared to GWGPR and GPR. The results show that the GWANN model results in eight groups using the three highest value of the variable importance.

MODELING NOVEL COVID-19 PANDEMIC IN NIGERIA USING COUNT DATA REGRESSION MODELS

This study aimed to model COVID-19 daily cases in Nigeria, focusing on confirmed, active, critical, recovered, and death cases using count data regression models. Three count data regression models-Poisson regression, Negative Binomial regression, and Generalized Poisson regression were applied to predict COVID-19 related deaths based on the mentioned variables. Secondary data from the Nigeria Centre for Disease Control (NCDC) between February 29, 2020, and October 19, 2020, were used. The study found that Poisson Regression could not handle over-dispersion inherent in the data. Consequently, Negative Binomial Regression and Generalized Poisson Regression were considered, with Generalized Poisson Regression identified as the best model through performance criteria such as -2 log likelihood (-2logL), Akaike information criterion (AIC), and Bayesian information criterion (BIC). The study revealed positive and significant impacts of confirmed, active, and critical cases on COVID-19 related deaths, while recovered cases had a negative effect. Recommendations included increased attention to confirmed, active, and critical cases by relevant authorities to mitigate COVID-19-related deaths in Nigeria.

RAINFALL MODELING USING THE GEOGRAPHICALLY WEIGHTED POISSON REGRESSION METHOD

Rainfall is an important parameter in understanding the climate and environment in the Malang Regency area. This research aims to model the distribution of rainfall in this region using the Geographically Weighted Poisson Regression (GWPR) method. GWPR is a spatial statistical approach that allows us to understand changes in inhomogeneous rainfall patterns throughout the Malang Regency area. Rainfall data collected from weather stations over several years was used in this study. We use GWR to study the relationship between various environmental factors, such as topography, vegetation, and land use, and rainfall distribution in Malang Regency. The results of the GWR analysis provide a deeper understanding of the spatial differences in the influence of these factors on rainfall. By applying GWR, we can find out how certain factors contribute to different rainfall patterns in certain regions. Rainfall modeling using the Geographically Weighted Poisson Regression (GWPR) method combines the power of Poisson regression in analyzing calculated data with the advantages of GWR in modeling spatial variability. GWPR allows us to identify and map rainfall distribution patterns that vary in geographic space. The main advantage of GWPR is its ability to provide local adjustments and capture the spatial variability associated with rainfall distribution. The results of the modeling analysis show that the GWPR is better, marked by the smallest AIC value, namely 336.84, compared to the generalized poisson regression model, namely 337.76.

A generalized Poisson regression analysis of determinants of early neonatal mortality in Ethiopia using 2019 Ethiopian mini demographic health survey

Neonatal mortality within the first few days of life is a pressing issue in sub-Saharan Africa, including Ethiopia. Despite efforts to achieve the targets set by the Sustainable Development Goals, the rate of neonatal mortality in Ethiopia has increased from 29 to 33 deaths per 1000 live births. This study aimed to investigate and identify significant determinants of neonatal mortality within the first 72 h of life in Ethiopia. Utilizing data from the 2019 Ethiopia Demographic and Health Survey, we employed Generalized Poisson regression analysis following rigorous model fitness assessment. Our study encompassed 5527 weighted live-born neonates. Among women in their reproductive years, 3.1% (n = 174) experienced at least one very early neonatal death. Multiple births (Incidence Risk Ratio (IRR) = 3.48; CI = 1.76, 6.887) and birth order six or above (IRR = 2.23; CI = 1.008, 4.916) were associated with an increased risk of neonatal death within the first 72 h. Conversely, household size (IRR = 0.72; CI = 0.586, 0.885) and additional feeding practices (IRR = 0.33; CI = 0.188, 0.579) were found to mitigate the risk of very early neonatal mortality per mother in Ethiopia. Interventions targeting the identified risk factors and promoting protective factors can contribute to reducing very early neonatal mortality rates and improving the well-being of mothers and their newborns. Further research and implementation of evidence-based strategies are needed to address these challenges and ensure better neonatal outcomes in Ethiopia.

The Applications of Generalized Poisson Regression Models to Insurance Claim Data

Predictive modeling has been widely used for insurance rate making. In this paper, we focus on insurance claim count data and address their common issues with more flexible modeling techniques. In particular, we study the zero-inflated and hurdle-generalized Poisson and negative binomial distributions in a functional form for modeling insurance claim count data. It is shown that these models are useful in addressing the problem of excess zeros and over-dispersion of the claim count variable. In addition, we show that including the exposure as a covariate in both the zero and the count part of the model is an effective approach to incorporating exposure information in zero-inflated and hurdle models. We illustrate the effectiveness and versatility of the introduced models using three real datasets. The results suggest their promising applications in insurance risk classification and beyond.

Examining the effect of vehicle type on right-turn crossing conflicts of minor road traffic at unsignalized T-intersections

In India, the crossing conflicts between right-turn vehicles of minor road traffic and conflicting vehicles on major roads have become more severe at unsignalized T-intersections in recent times. Due to the rapid growth of vehicular traffic, including motorized two-wheelers, auto-rickshaws, etc., results in an increase in right-turn crossing conflicts (RTCC) at unsignalized T-intersections. The severity of right-turn vehicles is related to the characteristics of both right-turning and conflicting through vehicles. Therefore, this study examines the effect of vehicle types on RTCC. The RTCC are observed using time-to-collision (TTC), at five unsignalized T-intersections in three different cities in India. The RTCC are categorized into critical RTCC and non-critical RTCC based on TTC threshold values obtained by the k-means clustering algorithm. A Generalized Poisson Regression model was developed using Python software. The study results revealed that the presence of higher composition of two-wheelers, auto-rickshaws, and cars in right-turn and conflicting through vehicles significantly influences the severity of RTCC. Also, the model results concluded that the speeds of right-turn and conflicting through vehicles, conflicting through traffic, right-turn traffic, vehicle gap, waiting time, and abnormal driving paths significantly affect the RTCC at unsignalized T-intersections. The findings of this study help traffic engineers and safety experts identify the critical unsignalized T-intersections using the number of right-turn crossing conflicts.

Testing latent classes in gut microbiome data using generalized Poisson regression models.

Human microbiome research has gained increasing importance due to its critical roles in comprehending human health and disease. Within the realm of microbiome research, the data generated often involves operational taxonomic unit counts, which can frequently present challenges such as over-dispersion and zero-inflation. To address dispersion-related concerns, the generalized Poisson model offers a flexible solution, effectively handling data characterized by over-dispersion, equi-dispersion, and under-dispersion. Furthermore, the realm of zero-inflated generalized Poisson models provides a strategic avenue to simultaneously tackle both over-dispersion and zero-inflation. The phenomenon of zero-inflation frequently stems from the heterogeneous nature of study populations. It emerges when specific microbial taxa fail to thrive in the microbial community of certain subjects, consequently resulting in a consistent count of zeros for these individuals. This subset of subjects represents a latent class, where their zeros originate from the genuine absence of the microbial taxa. In this paper, we introduce a novel testing methodology designed to uncover such latent classes within generalized Poisson regression models. We establish a closed-form test statistic and deduce its asymptotic distribution based on estimating equations. To assess its efficacy, we conduct an extensive array of simulation studies, and further apply the test to detect latent classes in human gut microbiome data from the Bogalusa Heart Study.

Application of Generalized Poisson Regression for the development of modeling motorcycle accidents based on rider characteristics

Motorcycles are a mode of transportation that has experienced many traffic accidents in developing countries. This incident can occur intentionally or not and cause great loss of life and property. Therefore, this work aims to develop a model of the number of motorcycle accidents based on driver characteristics using Generalized Poisson Regression in East Java Province, Indonesia. The modeling can identify the relationship between the number of motorcycle accidents with variables that contribute to accident based on driver characteristics. This characteristic is a combination of socioeconomic factors and movements. Research analysis includes descriptive statistics and accident modeling using Generalized Poisson Regression Models. Model parameter values are estimated using the Maximum Likelihood Estimation (MLE) approach. The optimal solution for parameter estimation uses the Newton-Raphson algorithm. The results found two models and then selected the best model. Based on the parameter significance test, the population density, the percentage of low education, the gender ratio, and the percentage of accident perpetrators without a driver’s license significantly impact the number of motorcycle accidents in East Java Province.

AGE AT MIGRATION, DURATION OF RESIDENCE, AND FERTILITY AMONG FEMALE MIGRANTS INTO JAVA

Java Island has become main destination for inter-regional migrants in Indonesia for decades. Previous studies has shown that migration can be expected to have a strong effect on fertility behavior of female migrants. The objective of this study is to examine fertility patterns of internal female migrants into Java Island. The generalized poisson regression models are used to investigate the socialization, the disruption, and the adaptation hypotheses in explaining the fertility differences between migrants and non-migrants. This study finds that the fertility rate for migrants is higher relative to the non-migrants. The analysis shows that first, there is a fertility assimilation process in migration into Java Island with fertility increasing gradually with age at migration. Second, there is a fertility disruption process in migration into Java Island with short-term fertility-lowering-effect of migration event. However, this study find no evidence on fertility adaptation process in migration into Java Island.

Using Spatial Modeling for Identifying Determinants Influencing Crash Estimate: Case Study of Hamedan, Iran

Road crashes are a major cause of fatalities worldwide, posing significant challenges for road-safety experts in selecting appropriate crash-frequency estimation models. This study introduces localized safety performance functions (C-SPFs), which explore the spatial variation of crash frequency and the spatial correlation between dependent variables. The exploratory spatial regression method is employed to identify optimal spatial associations. The study further predicts crashes using geographically weighted Poisson regression (GWPR) and generalized Poisson regression. Results indicate that C-SPFs offer greater accuracy than do models calibrated solely on annual average daily traffic. Moreover, the proposed model is especially relevant for jurisdictions facing higher heavy-vehicle traffic and frequent crashes. The development of C-SPFs and the use of GWPR provide valuable tools for policymakers and road-safety experts in enhancing crash-frequency estimation accuracy. Implementing these techniques can aid in prioritizing safety measures and countermeasures, especially in regions with significant heavy-vehicle traffic and crash occurrences. Additionally, the integration of spatial-analysis techniques and localized models can lead to more effective transportation planning and targeted road-safety interventions, ultimately contributing to reducing the burden of road crashes on a global scale.

POISSON REGRESSION MODELS TO ANALYZE FACTORS THAT INFLUENCE THE NUMBER OF TUBERCULOSIS CASES IN JAVA

Tuberculosis is an infectious disease and one of the world's top 10 highest causes of mortality in Indonesia. Based on this fact, it is necessary to study what factors affect number of tuberculosis cases. The number of tuberculosis cases as dependent variable is a count data that generally analyzed using Poisson regression. However, equidispersion assumption must be met, so Generalized Poisson Regression and Negative Binomial Regression are applied if the assumption is not met. Spatial aspects can be considered so Geographically Weighted Generalized Poisson Regression and Geographically Weighted Negative Binomial Regression were also conducted. Four models were built to evaluate relationship between number of tuberculosis cases and factors affecting it in Java in 2020. The explanatory variables are population density, percentage of children receiving BCG immunization, percentage of poor people, percentage of eligible drinking water facilities, percentage of family cards with access to proper sanitation, percentage of public places meet health requirements, and percentage of food management places meet hygienic requirements. This study shows that the best model for modeling the data is GWNBR with 2 groups of significant explanatory variables. Seven explanatory variables are statistically significant in 88 districts and six explanatory variables statistically significant in 12 districts.

Modeling the Number of Pneumonia in Toddlers in East Java Province in 2021 with Generalized Poisson Regression

Modeling the Number of Pneumonia in Toddlers in East Java Province in 2021 with Generalized Poisson Regression

Association Between Bowel Movement Pattern and Cognitive Function: Prospective Cohort Study and a Metagenomic Analysis of the Gut Microbiome.

Little is known regarding the association between intestinal motility patterns and cognitive function in individuals who are baseline cognitively healthy. The gut microbiome may contribute to the association. We examined the association between bowel movement (BM) pattern and cognitive function and explored the role of the gut microbiome in explaining this association. In this prospective study, we leveraged 3 cohort studies, Nurses' Health Study (NHS), NHSII, and Health Professionals Follow-Up Study (HPFS). Participants reported BM frequency and subjective cognitive function. In a subset of NHSII participants, we assessed cognitive function using an objective neuropsychological battery. We profiled the gut microbiome in a subset of participants using whole-genome shotgun metagenomics. General linear models, Poisson regression, and logistic regression were used to quantify the association of BM frequency with different cognitive measurements. We followed 112,753 men and women (women: 87.6%) with a mean age of 67.2 years at baseline (NHS: 76 years, NHSII: 59 years, HPFS: 75 years) for a median follow-up of 4 years (NHSII and HPFS: 4 years, NHS: 2 years). Compared with those with BM once daily, participants with BM frequency every 3+ days had significantly worse objective cognitive function, equivalent to 3.0 (95% confidence interval [CI],1.2-4.7) years of chronological cognitive aging. We observed similar J-shape dose-response relationships of BM frequency with the odds of subjective cognitive decline and the likelihood of having more subsequent subjective cognitive complaints (both p nonlinearity < 0.001). BM frequencies of every 3+ days and ≥twice/day, compared with once daily, were associated with the odds ratios of subjective cognitive decline of 1.73 (95% CI 1.60-1.86) and 1.37 (95% CI 1.33-1.44), respectively. BM frequency and subjective cognitive decline were significantly associated with the overall gut microbiome configuration (both p < 0.005) and specific microbial species in the 515 participants with microbiome data. Butyrate-producing microbial species were depleted in those with less frequent BM and worse cognition, whereas a higher abundance of proinflammatory species was associated with BM frequency of ≥twice/day and worse cognition. Lower BM frequency was associated with worse cognitive function. The gut microbial dysbiosis may be a mechanistic link underlying the association.