Geographically weighted temporally correlated logistic regression model

Copula-based markov chain logistic regression modeling on binomial time series data

A first-order autoregressive time series model with binomial distributed random variables has been developed using the copula-based Markov chain model approach. By still utilizing conditional probability, covariate variables can also be included in the model and can be assumed as the independent variable. The time series dependent variable with a binomial distribution and continuous independent variables can be modelled using a copula-based Markov chain model with the probability of success expressed in the logit model. This study proposes a copula-based Markov chain logistic regression model with marginal binomial and joint distribution functions built through the copula function. Besides that, this study aims to estimate the parameters involved in the model. The parameters are the parameters of the logistic regression model as the relationship between the dependent and independent variables and the copula parameter as a time dependency. Using the bivariate copula functions are Clayton, Gumbel and Frank, the parameter estimation method is Maximum Likelihood Estimation (MLE). Simulations were carried out to see the efficiency of the parameter estimation and asymptotic results. Based on the simulation results, it was concluded that MLE provides an accurate estimate of the copula-based Markov chain logistic regression. In addition, the copula-based Markov chain logistic regression model can not only see the relationship between the independent and dependent variables but also provide an estimate of the time dependency of the dependent variable. The following are some of the proposed approach's highlights:•This method proposes a binomial time series data model with covariate variables by combining the logistic regression model and the first-order Markov chain model.•Parameter estimation in this model uses the Maximum Likelihood Estimation method.•The model provides the possibility to see the relationship between variables and the time dependency.

How do drunk-driving events escalate into drunk-driving crashes? An empirical analysis of Beijing from a spatiotemporal perspective

Drunk-driving events often escalate into drunk-driving crashes, however, the contributing factors of this progression remain elusive. To mitigate the likelihood of crashes stemming from drunk-driving events, this paper introduces the notion of 'the severity of drunk-driving event' and examines the complex relationship between the severity and its contributing factors, considering spatiotemporal heterogeneity. The study utilizes a Geographically and Temporally Weighted Binary Logistic Regression (GTWBLR) model to conduct spatiotemporal analysis based on police-reported drunk-driving events in Beijing, China. The results show that most factors passed the non-stationary test, indicating their effects on the severity of drunk-driving event vary significantly across different spatial and temporal domains. Notably, during non-workday, drunk-driving events in northeast of Beijing are more likely to escalate into crashes. Furthermore, severe weather during winter in the northwest of Beijing is associated with high risk of drunk-driving crashes. Based on these insights, the authorities can strengthen drunk-driving checks in the northeast region of Beijing, particularly during non-workdays. And it is crucial to promptly clear accumulated snow on the roads during severe winter weather to improve road safety. These insights and recommendations are highly valuable for reducing the risk of drunk-driving crashes.

Spatial multilevel analysis of age at death of under-5 children and associated determinants: EDHS 2000-2016

OBJECTIVE

This study examines trends, spatial distribution and determinants of age at death of under-5 children in Ethiopia.

DESIGN

This study used secondary data from the 2000, 2005, 2011 and 2016 Ethiopian Demographic and Health Surveys. A multilevel partial ordinal logistic regression model was used to assess the effects of variables on the age at death of children under 5 years.

SETTING

Ethiopia.

PARTICIPANTS

The final analysis included a sample of 3997 deaths of newborns, infants and toddlers.

RESULTS

A total of 1508, 1054, 830 and 605 deaths of under-5 children were recorded in the 2000, 2005, 2011 and 2016 survey years, respectively. The death of newborns, infants and toddlers showed a significant decrease from 2000 to 2016, with reductions of 33.3% to 17.4%, 42.4% to 12.6% and 45.2% to 11.6%, respectively. The analysis using Global Moran's Index revealed significant spatial autocorrelation in mortality for each survey year (p<0.05). The intraclass correlation of age at death of under-5 children within regions was substantial. Furthermore, the odds of newborn deaths among under-5 children (OR: 0.638, 95% CI: 0.535, 0.759) were lower for those delivered in health institutions compared with those delivered at home.

CONCLUSIONS

Throughout the survey years from 2000 to 2016, newborn children had higher mortality rates compared with infants and toddlers, and significant spatial variations were observed across different zones in Ethiopia. Factors such as child's sex, age of mother, religion, birth size, sex of household head, place of delivery, birth type, antenatal care, wealth index, spatial autocovariate, Demographic and Health Survey year, place of residence and region were found to be significant in influencing the death of under-5 children in Ethiopia. Overall, there has been a decreasing trend in the proportion of under-5 child mortality over the four survey years in Ethiopia.

CLUMondo-BNU for simulating land system changes based on many-to-many demand-supply relationships with adaptive conversion orders

Land resources are fundamentally important to human society, and their transition from one macroscopic state to another is a vital driving force of environment and climate change locally and globally. Thus, many efforts have been devoted to the simulations of land changes. Among all spatially explicit simulation models, CLUMondo is the only one that simulates land changes by incorporating the multifunctionality of a land system and allows the establishment of many-to-many demand-supply relationships. In this study, we first investigated the source code of CLUMondo, providing a complete, detailed mechanism of this model. We found that the featured function of CLUMondo-balancing demands and supplies in a many-to-many mode-relies on a parameter called conversion order. The setting of this parameter is a manual process and requires expert knowledge, which is not feasible for users without an understanding of the whole, detailed mechanism. Therefore, the second contribution of this study is the development of an automatic method for adaptively determining conversion orders. Comparative experiments demonstrated the validity and effectiveness of the proposed automated method. We revised the source code of CLUMondo to incorporate the proposed automated method, resulting in CLUMondo-BNU v1.0. This study facilitates the application of CLUMondo and helps to exploit its full potential.

Generalized Geographically Weighted Regression Model within a Modularized Bayesian Framework

Geographically weighted regression (GWR) models handle geographical dependence through a spatially varying coefficient model and have been widely used in applied science, but its general Bayesian extension is unclear because it involves a weighted log-likelihood which does not imply a probability distribution on data. We present a Bayesian GWR model and show that its essence is dealing with partial misspecification of the model. Current modularized Bayesian inference models accommodate partial misspecification from a single component of the model. We extend these models to handle partial misspecification in more than one component of the model, as required for our Bayesian GWR model. Information from the various spatial locations is manipulated via a geographically weighted kernel and the optimal manipulation is chosen according to a Kullback-Leibler (KL) divergence. We justify the model via an information risk minimization approach and show the consistency of the proposed estimator in terms of a geographically weighted KL divergence.

Stochastic approximation cut algorithm for inference in modularized Bayesian models

Bayesian modelling enables us to accommodate complex forms of data and make a comprehensive inference, but the effect of partial misspecification of the model is a concern. One approach in this setting is to modularize the model and prevent feedback from suspect modules, using a cut model. After observing data, this leads to the cut distribution which normally does not have a closed form. Previous studies have proposed algorithms to sample from this distribution, but these algorithms have unclear theoretical convergence properties. To address this, we propose a new algorithm called the stochastic approximation cut (SACut) algorithm as an alternative. The algorithm is divided into two parallel chains. The main chain targets an approximation to the cut distribution; the auxiliary chain is used to form an adaptive proposal distribution for the main chain. We prove convergence of the samples drawn by the proposed algorithm and present the exact limit. Although SACut is biased, since the main chain does not target the exact cut distribution, we prove this bias can be reduced geometrically by increasing a user-chosen tuning parameter. In addition, parallel computing can be easily adopted for SACut, which greatly reduces computation time.

Examining the spatially varying effects of factors on PM2.5 concentrations in Chinese cities using geographically weighted regression modeling

Whilst numerous studies have explored the spatial patterns and underlying causes of PM_2.5, little attention has been paid to the spatial heterogeneity of the factors affecting PM_2.5. In this study, a geographically weighted regression (GWR) model was used to explore the strength and direction of nexus between various factors and PM_2.5 in Chinese cities. A comprehensive interpretive framework was established, composed of 18 determinants spanning the three categories of natural conditions, socioeconomic factors, and city features. Our results indicate that PM_2.5 concentration levels were spatially heterogeneous and markedly higher in cities in eastern China than in cities in the west of the country. Based on the results of GWR, significant spatial heterogeneity was identified in both the direction and strength of the determinants at the local scale. Among all of the natural variables, elevation was found to be statistically significant with its effects on PM_2.5 in 95.60% of the cities and it correlated negatively with PM_2.5 in 99.63% cities, with its effect gradually weakening from the eastern to the western parts of China. The variable of built-up areas emerged as the strongest variable amongst the socioeconomic variables studied; it maintained a positive significant relationship in cities located in the Pearl River Delta and surrounding areas, while in other cities it exhibited a negative relationship to PM_2.5. The highest coefficients were located in cities in northeast China. As the strongest variable amongst the six landscape factors, patch density maintained a positive relationship in part of cities. While in cities in the northeast regions, patch density exhibited a negative relationship with PM_2.5, revealing that increasing urban fragmentation was conducive to PM_2.5 reductions in those regions. These empirical results provide a basis for the formulation of targeted and differentiated air quality improvement measures in the task of regional PM_2.5 governances.

Spatiotemporal Influence of Urban Environment on Taxi Ridership Using Geographically and Temporally Weighted Regression

Taxicabs play an important role in urban transit systems, and their ridership is significantly influenced by the urban built environment. The intricate relationship between taxi ridership and the urban environment has been explored using either conventional ordinary least squares (OLS) regression or geographically weighted regression (GWR). However, time constitutes a significant dimension, particularly when analyzing spatiotemporal hourly taxi ridership, which is not effectively incorporated into conventional models. In this study, the geographically and temporally weighted regression (GTWR) model was applied to model the spatiotemporal heterogeneity of hourly taxi ridership, and visualize the spatial and temporal coefficient variations. To test the performance of the GTWR model, an empirical study was implemented for Xiamen city in China using a set of weekday taxi pickup point data. Using point-of-interest (POI) data, hourly taxi ridership was analyzed by incorporating it to various spatially urban environment variables based on a 500 × 500 m grid unit. Compared to the OLS and GWR, the GTWR model obtained the best performance, both in terms of model fit and explanatory accuracy. Moreover, the urban environment was revealed to have a significant impact on taxi ridership. Road density was found to decrease the number of taxi trips in particular places, and the density of bus stops competed with taxi ridership over time. The GTWR modelling provides valuable insights for investigating taxi ridership variation as a function of spatiotemporal urban environment variables, thereby facilitating an optimal allocation of taxi resources and transportation planning.