• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.17 no.4
• /
• pp.14-25
• /
• 2018

In recent years, it has become technically easier to explain factors related with traffic accidents in the Big Data era. Therefore, it is necessary to apply the latest analysis techniques to analyze the traffic accident data and to seek for new findings. The purpose of this study is to compare the predictive performance of the negative binomial regression model and the deep learning method developed in this study to predict the frequency of traffic accidents in expressways. As a result, the MOEs of the deep learning model are somewhat superior to those of the negative binomial regression model in terms of prediction performance. However, using a deep learning model could increase the predictive reliability. However, it is easy to add other independent variables when using deep learning, and it can be expected to increase the predictive reliability even if the model structure is changed.

• Journal of Korean Society of Transportation
• /
• v.25 no.3
• /
• pp.123-135
• /
• 2007

The objective of this study is to analyze the relationship between traffic accidents occurring at trumpet interchange ramps according to accident type as well as the relevant factors that led to the traffic accidents, such as geometric design elements and traffic volumes. In the process of analysis of the distribution of traffic accidents, negative binomial distribution was selected as the most appropriate model. Negative binomial regression models were developed for total trumpet interchange ramps, direct ramps, loop ramps and semi-direct ramps based on the negative binomial distribution. Based upon several statistical diagnostics of the difference between observed accidents and predicted accidents with four previously developed models, the fit proved to be reasonable. Understanding of statistically significant variables in the developed model will enable designers to increase efficiency in terms of road operations and the development of traffic accident prevention policies in accordance with road design features.

• Journal of Korean Society of Transportation
• /
• v.23 no.3 s.81
• /
• pp.125-136
• /
• 2005

The current traffic accident reduction procedure in economic feasibility study does not consider the characteristics of road and V/C ratio. For solving this problem, this paper suggests methods to be able to evaluate safety of each road in construction and improvement through developing accident Prediction model in reflecting V/C ratio Per road types and traffic characters. In this paper as primary process, model is made by tke object of urban roads. Most of all, factor effecting on accident relying on road types is selected. At this point, selecting criteria chooses data obtained from road planning procedure, traffic volume, existence or non-existence of median barrier, and the number of crossing point, of connecting road. and of traffic signals. As a result of analyzing between each factor and accident. all appear to have relatives at a significant level of statistics. In this research, models are classified as 4-categorized classes according to roads and V/C ratio and each of models draws accident predicting model through Poisson regression along with verifying real situation data. The results of verifying models come out relatively satisfactory estimation against real traffic data. In this paper, traffic accident prediction is possible caused by road's physical characters by developing accident predicting model per road types resulted in V/C ratio and this result is inferred to be used on predicting accident cost when road construction and improvement are performed. Because data using this paper are limited in only province of Jeollabuk-Do, this paper has a limitation of revealing standards of all regions (nation).

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.22 no.1
• /
• pp.1-15
• /
• 2023

In this study, the number of truck traffic accidents was predicted by using Poisson and negative binomial regression analysis to understand what factors affect accidents using expressway data. Significant variables in the truck traffic accident prediction model were continuous driving time, link length, truck traffic volume. number of bridges and number of drowsy shelters. The calculated LOSS rating was expressed on the national expressway network to diagnose the risk of truck accidents. This is expected to be used as basic data for policy establishment to reduce truck accidents on expressways.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.15 no.6
• /
• pp.1123-1130
• /
• 2020

This study aims to classify the severity in car crashes using five classification learning models. The dataset used in this study contains 21,013 vehicle crashes, obtained from Korea Expressway Corporation, between the year of 2015-2017 and the LightGBM(Light Gradient Boosting Model) performed well with the highest accuracy. LightGBM, the number of involved vehicles, type of accident, incident location, incident lane type, types of accidents, types of vehicles involved in accidents were shown as priority factors. Based on the results of this model, the establishment of a management strategy for response of highway traffic accident should be presented through a consistent prediction process of accident severity level. This study identifies applicability of Machine Learning Models for Predicting of the Severity of Car Traffic Accidents on a Highway and suggests that various machine learning techniques based on big data that can be used in the future.

• Journal of Korean Society of Transportation
• /
• v.32 no.4
• /
• pp.369-379
• /
• 2014

The Statistical regression model has been used to construct crash prediction models, despite its limitations in assuming data distribution and functional form. In response to the limitations associated with the statistical regression models, a few studies based on non-parametric methods such as neural networks have been proposed to develop crash prediction models. However, these models have a major limitation in that they work as black boxes, and therefore cannot be directly used to identify the relationships between crash frequency and crash factors. A genetic programming model can find a solution to a problem without any specified assumptions and remove the black box effect. Hence, this paper investigates the application of the genetic programming technique to develope the crash prediction model. The data collected from the Gyeongbu expressway during the past three years (2010-2012), were separated into straight and curve sections. The random forest technique was applied to select the important variables that affect crash occurrence. The genetic programming model was developed based on the variables that were selected by the random forest. To test the goodness of fit of the genetic programming model, the RMSE of each model was compared to that of the negative binomial regression model. The test results indicate that the goodness of fit of the genetic programming models is superior to that of the negative binomial models.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2020.06a
• /
• pp.301-301
• /
• 2020

물은 보다 나은 삶을 살아가는데 있어 매우 중요한 자원이며, 안전하고 깨끗한 물을 공급받는 것은 국민 생활 영위에 반드시 필요한 부분이다. 국내 주요 수자원은 하천수를 통해 확보하고 있으며, 안전한 수자원 공급을 위해서는 하천관리를 통한 수질오염사고에 대비한 대책 수립이 필요하다. 국내에서는 페놀, 황산 등 독성오염물질 유출로 인한 수질오염사고가 발생한 바 있고, 그 피해액이 수백억에 달한다. 이러한 수질오염사고로 인한 피해액을 감소시키고 안전한 수자원 공급을 유지하기 위해서는 오염물질의 거동을 이해하고 예측하는 것이 매우 중요하다. 국내하천의 경우, 대부분 하폭 대비 수심비가 크기 때문에 오염물질이 2차원 혼합특성을 나타낸다. 따라서 본 연구에서는 하천 내 오염물질의 2차원적 혼합거동을 해석할 수 있는 수치모형을 개발하고자 하며, 현장에 적합한 해석기법을 검토하고 모형 개발 방향을 결정하고자 한다. 본 연구에서는 하천 내 수질오염사고 발생 시 신속하고 정확한 수질 분석 및 예측을 목표로 오염물질 혼합해석에 주로 활용되는 격자기반 모형과 입자추적 기반 모형의 프로토타입을 개발했다. 용존성 오염물질을 대상으로 격자 기반 및 입자 기반 혼합해석 모형을 개발했으며, 오염물질의 주입형태와 하천 내 유속 분포를 가정해 혼합해석을 수행했다. 격자 기반 모형의 경우, 경계조건과 분산계수의 결정이 필요하고 수렴/발산 문제로 인해 모형의 안정적 실행을 위한 조건 수립이 필요하다. 입자 기반 모형의 경우에도 입자 수에 따른 계산시간 개선이 필요하지만, 입력조건 결정이 간편하고 분산계수 입력이 필요 없어 신속한 모의조건 설정이 가능하다. 오염물질 혼합해석 모형 개발을 위한 해석기법 검토 결과, 신속한 수질 분석 및 예측 결과를 제공하기 위해서는 계산시간 개선을 전제로 모의조건 설정이 용이한 입자 기반 모형이 가장 적합한 것으로 판단된다.

• Journal of Korean Society of Transportation
• /
• v.30 no.2
• /
• pp.101-116
• /
• 2012

Accident prediction models have been utilized to predict accident possibilities in existing or projected freeways and to evaluate programs or policies for improving safety. In this study, a traffic accident prediction model for freeways was developed for the above purposes. When selecting variables for the model, the highest priority was on the ease of both collecting data and applying them into the model. The dependent variable was set as the number of total accidents and the number of accidents including casualties in the unit of IC(or JCT). As a result, two models were developed; the overall accident model and the casualty-related accident model. The error structure adjusted to each model was the negative binomial distribution and the Poisson distribution, respectively. Among the two models, a more appropriate model was selected by statistical estimation. Major nine national freeways were selected and five-year dada of 2003~2007 were utilized. Explanatory variables should take on either a predictable value such as traffic volumes or a fixed value with respect to geometric conditions. As a result of the Maximum Likelihood estimation, significant variables of the overall accident model were found to be the link length between ICs(or JCTs), the daily volumes(AADT), and the ratio of bus volume to the number of curved segments between ICs(or JCTs). For the casualty-related accident model, the link length between ICs(or JCTs), the daily volumes(AADT), and the ratio of bus volumes had a significant impact on the accident. The likelihood ratio test was conducted to verify the spatial and temporal transferability for estimated parameters of each model. It was found that the overall accident model could be transferred only to the road with four or more than six lanes. On the other hand, the casualty-related accident model was transferrable to every road and every time period. In conclusion, the model developed in this study was able to be extended to various applications to establish future plans and evaluate policies.

• Journal of Korean Society of Transportation
• /
• v.31 no.1
• /
• pp.65-76
• /
• 2013

Traffic accidents at signalized intersections have been increased annually so that it is required to examine the causation to reduce the accidents. However, the current existing accident models were developed mainly by using non-linear regression models such as Poisson methods. These non-linear regression methods lack to reveal the complicated causation for traffic accidents, though they are the right choice to study randomness and non-linearity of accidents. Therefore, it is required to utilize another statistical method to make up for the lack of the non-linear regression methods. This study developed accident prediction models for 4 legged signalized intersections with Poisson methods and compared them with structural equation models. This study used structural equation methods to reveal the complicated causation of traffic accidents, because the structural equation method has merits to explain more causational factors for accidents than others.

• Journal of Korean Society of Transportation
• /
• v.27 no.3
• /
• pp.91-102
• /
• 2009

A traffic accident prediction model developed using various design variables(road design variables, geometric variables, and traffic environmental variables) is one of the most important factors to safety design evaluation system for roads. However, statistical accident models have a crucial problem not applicable for all intersections. To make up this problem, this study developed AMFs(Accident Modification Factors) through statistical modeling methods, historical accident databases, judgment from traffic experts, and literature review by considering design variable's characteristics, traffic accident rates, and traffic accident frequency. AMFs developed in this study include exclusive left-turn lane, exclusive right-turn lane, sight distance, and intersection angle. Predictabilities of the developed AMFs and the existing accident prediction models are compared with real accident historical data. The results showed that performances of the developed AMFs are superior to the existing statistical accident prediction models. These findings show that AMFs should be considered as a important process to develop safety design evaluation algorithms. Additionally, AMFs could be used as an index that can judge the impact of corresponding design variables on accidents in rural intersections.