• Journal of the Korean Society of Safety
• /
• v.29 no.5
• /
• pp.154-159
• /
• 2014

This study deals with the pedestrian accidents of intersections in case of Cheongju. The objective is to develop the pedestrian accident models using Tobit regression model. In pursuing the above, the pedestrian accident data from 2007 to 2011 were collected from TAAS data set of Road Traffic Authority. To analyze the accident, Poisson, negative binomial and Tobit regression models were utilized in this study. The dependent variable were the number of accident by intersection. Independent variables are traffic volume, intersection geometric structure and the transportation facility. The main results were as follows. First, Tobit model was judged to be more appropriate model than other models. Also, these models were analyzed to be statistically significant. Second, such the main variables related to accidents as traffic volume, pedestrian volume, number of traffic island, crossing length and the pedestrian countdown signal systems were adopted in the above model.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2020.06a
• /
• pp.208-208
• /
• 2020

지구온난화와 하천환경의 변화로 수질 오염이 심각해지고 녹조 현상 등의 피해가 증가하고 있다. 특히, 기후변화로 인해 온도와 강우량의 변동성이 동시에 증가하고 있어 하천의 수환경 관리측면에서 어려움이 증가하고 있다. 최근 하천 개발 사업으로 인한 인공 구조물 축조로 하천의 오염도 변화는 중요한 논점으로 대두되었으며, 그에 따라 정확한 수질 전망이 요구되고 있다. 녹조평가에 있어 주요 대리변수(proxy variable)로 chlorophyll-a(Chl-a)가 사용되고 있지만, Chl-a는 규조류와 남조류(cyanobacteria) 모두에서 발견되는 지표로서, 녹조의 유해성을 Chl-a 수질 지표만을 사용하여 판단하기에는 한계가 있다. Chl-a뿐만 아니라 수질에 대한 유량, 온도, 영양염류 등의 영향 또한 기존 연구에서 밝혀진 바 있다. 하지만 기존의 물리기반의 결정론적모형은 수질의 추계학적(stochastic) 특성을 반영하는데 제한적이며, 다양한 수문기상학적 조건을 고려한 시나리오 기반의 분석을 수행하는데 한계가 있다. 따라서 본 연구에서는 특정 지점의 보 건설 이후 수문기상 자료를 이용하여 유해 남조류 개체수와 관계있는 수문기상학적 요인을 평가하고 최종적으로 Bayesian Poisson Regression 기반의 중·장기 녹조 예측 모형을 개발하였으며, 해설결과에 대한 불확실성 정보도 제공할 수 있도록 하였다.

• Journal of Preventive Medicine and Public Health
• /
• v.43 no.6
• /
• pp.543-548
• /
• 2010

Objectives: The objective of this study was to calculate sample size and power in an ongoing cohort, Korea radiation effect and epidemiology cohort (KREEC). Method: Sample size calculation was performed using PASS 2002 based on Cox regression and Poisson regression models. Person-year was calculated by using data from '1993-1997 Total cancer incidence by sex and age, Seoul' and Korean statistical informative service. Results: With the assumption of relative risk=1.3, exposure:non-exposure=1:2 and power=0.8, sample size calculation was 405 events based on a Cox regression model. When the relative risk was assumed to be 1.5 then number of events was 170. Based on a Poisson regression model, relative risk=1.3, exposure:non-exposure=1:2 and power=0.8 rendered 385 events. Relative risk of 1.5 resulted in a total of 157 events. We calculated person-years (PY) with event numbers and cancer incidence rate in the nonexposure group. Based on a Cox regression model, with relative risk=1.3, exposure:non-exposure=1:2 and power=0.8, 136 245PY was needed to secure the power. In a Poisson regression model, with relative risk=1.3, exposure:non-exposure=1:2 and power=0.8, person-year needed was 129517PY. A total of 1939 cases were identified in KREEC until December 2007. Conclusions: A retrospective power calculation in an ongoing study might be biased by the data. Prospective power calculation should be carried out based on various assumptions prior to the study.

• Communications for Statistical Applications and Methods
• /
• v.24 no.1
• /
• pp.43-65
• /
• 2017

We propose a flexible cure rate survival model by assuming that the number of competing causes of the event of interest has the Poisson distribution and the time for the event follows the gamma-G family of distributions. The extended family of gamma-G failure-time models with long-term survivors is flexible enough to include many commonly used failure-time distributions as special cases. We consider a frequentist analysis for parameter estimation and derive appropriate matrices to assess local influence on the parameters. Further, various simulations are performed for different parameter settings, sample sizes and censoring percentages. We illustrate the performance of the proposed regression model by means of a data set from the medical area (gastric cancer).

• Communications for Statistical Applications and Methods
• /
• v.11 no.2
• /
• pp.381-397
• /
• 2004

The marginal likelihood has become an important tool for model selection in Bayesian analysis because it can be used to rank the models. We discuss the marginal likelihood for Poisson regression models that are potentially useful in small area estimation. Computation in these models is intensive and it requires an implementation of Markov chain Monte Carlo (MCMC) methods. Using importance sampling and multivariate density estimation, we demonstrate a computation of the marginal likelihood through an output analysis from an MCMC sampler.

• Journal of the Korean Data and Information Science Society
• /
• v.26 no.6
• /
• pp.1513-1521
• /
• 2015

Selecting relevant variables for a statistical model is very important in regression analysis. Recently, variable selection methods using a penalized likelihood have been widely studied in various regression models. The main advantage of these methods is that they select important variables and estimate the regression coefficients of the covariates, simultaneously. In this paper, we propose a simple procedure based on a penalized h-likelihood (HL) for variable selection in Poisson hierarchical generalized linear models (HGLMs) for correlated count data. For this we consider three penalty functions (LASSO, SCAD and HL), and derive the corresponding variable-selection procedures. The proposed method is illustrated using a practical example.

• Journal of the Korean Data and Information Science Society
• /
• v.19 no.4
• /
• pp.1353-1360
• /
• 2008

An estimating procedure is introduced for the nonlinear mixed-effect Poisson regression, for longitudinal study, where data from different subjects are independent whereas data from same subject are correlated. The proposed procedure provides the estimates of the mean function of the response variables, where the canonical parameter is related to the input vector in a nonlinear form. The generalized cross validation function is introduced to choose optimal hyper-parameters in the procedure. Experimental results are then presented, which indicate the performance of the proposed estimating procedure.

• Communications for Statistical Applications and Methods
• /
• v.29 no.2
• /
• pp.239-250
• /
• 2022

In many applications, we frequently encounter correlated multiple outcomes measured on the same subject. Joint modeling of such multiple outcomes can improve efficiency of inference compared to independent modeling. For instance, in developmental toxicity studies, fetal weight and number of malformed pups are measured on the pregnant dams exposed to different levels of a toxic substance, in which the association between such outcomes should be taken into account in the model. The number of malformations may possibly have many zeros, which should be analyzed via zero-inflated count models. Motivated by applications in developmental toxicity studies, we propose a Bayesian joint modeling framework for continuous and count outcomes with excess zeros. In our model, zero-inflated Poisson (ZIP) regression model would be used to describe count data, and a subject-specific random effects would account for the correlation across the two outcomes. We implement a Bayesian approach using MCMC procedure with data augmentation method and adaptive rejection sampling. We apply our proposed model to dose-response analysis in a developmental toxicity study to estimate the benchmark dose in a risk assessment.

• International Journal of Highway Engineering
• /
• v.12 no.2
• /
• pp.43-49
• /
• 2010

This study deals with the traffic accident of the Cheongju arterial link sections. The purpose of the study is to develop the traffic accident model. In pursuing the above, this study gives particular attentions to developing the ZAM(zero-altered model) model using the accident data of arterial roads devided by 322 small link sections. The main results analyzed by ZIP(zero inflated Poisson model) and ZINB(zero inflated negative binomial model) which are the methods of ZAM, are as follows. First, the evaluation of various developed models by the Vuong statistic and t statistic for overdispersion parameter ${\alpha}$ shows that ZINB is analyzed to be optimal among Poisson, NB, ZIP(zero-inflated Poisson) and ZINB regression models. Second, ZINB is evaluated to be statistically significant in view of t, ${\rho}$ and ${\rho}^2$ (0.63) values compared to other models. Finally, the accident factors of ZINB models are developed to be the traffic volume(ADT), number of entry/exit and length of median. The traffic volume(ADT) and the number of entry/exit are evaluated to be the '+' factors and the length of median to be '-' factor of the accident.

• Journal of the Korean Data and Information Science Society
• /
• v.25 no.6
• /
• pp.1221-1229
• /
• 2014

In this paper we choose the best model among several bivariate Poisson models on Korean soccer data. The models considered allow for correlation between the number of goals of two competing teams. We use an R package called bivpois for bivariate Poisson regression models and the data of K-league for season 1983-2012. Finally we conclude that the best fitted model supported by the AIC and BIC is the bivariate Poisson model with constant covariance. The zero and diagonal inflated models did not improve the model fit. The model can be used to examine home-away effect, goodness of fit, attack and defense parameters.