Risk Factors Affecting Traffic Accidents at Urban Weaving Sections

[Full title: Risk Factors Affecting Traffic Accidents at Urban Weaving Sections: Evidence from China]

International Journal of Environmental Research and Public Health Article Risk Factors A ff ecting Tra ffi c Accidents at Urban Weaving Sections: Evidence from China Xinhua Mao 1,2, *, Changwei Yuan 1 , Jiahua Gan 3 and Shiqing Zhang 4 1 School of Economics and Management, Chang’an University, Xi’an 710064, China; changwei@chd.edu.cn 2 Department of Civil and Environmental Engineering, University of Waterloo, Waterloo, ON N 2 L 3 G 1, Canada 3 Transport Planning and Research Institute, Ministry of Transport, Beijing 100028, China; ganjh@tpri.org.cn 4 School of Management Engineering, Zhengzhou University of Aeronautics, Zhengzhou 450046, China; zshiqing_chd@163.com * Correspondence: mxinhua@uwaterloo.ca Received: 7 April 2019; Accepted: 29 April 2019; Published: 1 May 2019 Abstract: As a critical configuration of interchanges, the weaving section is inclined to be involved in more tra ffi c accidents, which may bring about severe casualties. To identify the factors associated with tra ffi c accidents at the weaving section, we employed the multinomial logistic regression approach to identify the correlation between six categories of risk factors (drivers’ attributes, weather conditions, tra ffi c characteristics, driving behavior, vehicle types and temporal-spatial distribution) and four types of tra ffi c accidents (rear-end, side wipe, collision with fixtures and rollover) based on 768 accident samples of an observed weaving section from 2016 to 2018. The modeling results show that drivers’ gender and age, weather condition, tra ffi c density, weaving ratio, vehicle speed, lane change behavior, private cars, season, time period, day of week and accident location are important factors a ff ecting tra ffi c accidents at the weaving section, but they have di ff erent contributions to the four tra ffi c accident types. The results also show that tra ffi c density of ≥ 31 vehicle / 100 m has the highest risk of causing rear-end accidents, weaving ration of ≥ 41% has the highest possibility to bring about a side wipe incident, collision with fixtures is the most likely to happen in snowy weather, and rollover is the most likely incident to occur in rainy weather Keywords: tra ffi c accidents; risk factors; weaving section; multinomial logistic regression 1. Introduction The weaving section is a common type of road configuration, which widely exists in freeway interchanges [ 1 ]. It is formed when a merging area is closely followed by a diverging area, typically within less than 0.76 km [ 2 ]. Based on the minimum number of lane changes required for completing the weaving behavior, weaving areas can be grouped into three major types, i.e., Type A, Type B, and Type C, shown in Figure 1 . Int. J. Environ. Res. Public Health 2019 , 16 , x; doi: www.mdpi.com/journal/ijerph Article Risk Factors Affecting Traffic Accidents at Urban Weaving Sections: Evidence from China Xinhua Mao 1,2, *, Changwei Yuan 1 , Jiahua Gan 3 and Shiqing Zhang 4 1 School of Economics and Management, Chang’an University, Xi’an 710064, China; changwei@chd.edu.cn 2 Department of Civil and Environmental Engineering, University of Waterloo, Waterloo, ON N 2 L 3 G 1, Canada 3 Transport Planning and Research Institute, Ministry of Transport, Beijing 100028, China; ganjh@tpri.org.cn 4 School of Management Engineering, Zhengzhou University of Aeronautics, Zhengzhou 450046, China; zshiqing_chd@163.com * Correspondence: mxinhua@uwaterloo.ca Received: 07 April 2019; Accepted: 29 April 2019; Published: date Abstract: As a critical configuration of interchanges, the weaving section is inclined to be involved in more traffic accidents, which may bring about severe casualties. To identify the factors associated with traffic accidents at the weaving section, we employed the multinomial logistic regression approach to identify the correlation between six categories of risk factors (drivers’ attributes, weather conditions, traffic characteristics, driving behavior, vehicle types and temporal-spatial distribution) and four types of traffic accidents (rear-end, side wipe, collision with fixtures and rollover) based on 768 accident samples of an observed weaving section from 2016 to 2018. The modeling results show that drivers’ gender and age, weather condition, traffic density, weaving ratio, vehicle speed, lane change behavior, private cars, season, time period, day of week and accident location are important factors affecting traffic accidents at the weaving section, but they have different contributions to the four traffic accident types. The results also show that traffic density of ≥ 31 vehicle/100 m has the highest risk of causing rear-end accidents, weaving ration of ≥ 41% has the highest possibility to bring about a side wipe incident, collision with fixtures is the most likely to happen in snowy weather, and rollover is the most likely incident to occur in rainy weather. Keywords: traffic accidents; risk factors; weaving section; multinomial logistic regression 1. Introduction The weaving section is a common type of road configuration, which widely exists in freeway interchanges [1]. It is formed when a merging area is closely followed by a diverging area, typically within less than 0.76 km [2]. Based on the minimum number of lane changes required for completing the weaving behavior, weaving areas can be grouped into three major types, i.e., Type A, Type B, and Type C, shown in Figure 1. ( a ) ( b ) ( c ) Figure 1 . Three types of weaving sections. ( a ) Type A weaving section; ( b ) Type B weaving section ; ( c ) Type C weaving section Figure 1. Three types of weaving sections. ( a ) Type A weaving section; ( b ) Type B weaving section; ( c ) Type C weaving section Int. J. Environ. Res. Public Health 2019 , 16 , 1542; doi:10.3390 / ijerph 16091542 www.mdpi.com / journal / ijerph

Int. J. Environ. Res. Public Health 2019 , 16 , 1542 2 of 17 The three types of weaving sections are defined in detail as follows [ 2 ]. • Type A weaving section: Every weaving vehicle has to make at least one lane change in the weaving area • Type B weaving section: One weaving movement can be made without making any lane change, while the other weaving movement requires at most one lane change • Type C weaving section: One weaving movement can be made without making any lane change, while the other weaving movement requires at least two lane changes In reality, it is also possible that two of these weaving section types can overlap [ 3 ]. Compared to regular road segments, tra ffi c flow characteristics at weaving sections are more complicated [ 4 ]. For example, in the weaving process, merging and diverging vehicles should enter their target lane by changing lanes in a limited distance at the weaving section without the aid of a tra ffi c control device [ 5 ]. Due to urban land constraints, an increasing number of interchanges have been built in China’s metropolitan areas, which greatly reduced travel time and improved tra ffi c capacity. However, as an important part of interchanges, weaving sections have become inclined to be involved in more tra ffi c accidents, such as rear-end and side wipe [ 6 ], which have brought about severe casualties and significant economic losses. To reduce tra ffi c conflicts at weaving sections and prevent them from becoming accident-prone locations, why and how tra ffi c crashes happen at weaving sections should be addressed Through wide-ranging literature resources, there are several kinds of research streams concentrating on tra ffi c accidents at weaving sections: (i) risk factor analysis of tra ffi c accidents [ 7 , 8 ], (ii) tra ffi c accident distribution characteristics [ 9 , 10 ], (iii) tra ffi c accident prediction [ 11 , 12 ], (iv) tra ffi c accident hazard point identification [ 13 , 14 ], (v) tra ffi c safety assessment [ 15 , 16 ] and (vi) tra ffi c accident prevention [ 17 , 18 ]. However, they failed to compare the e ff ects of various risk factors on di ff erent types of tra ffi c accidents at the weaving section and the correlation between tra ffi c accident types and the accident location at weaving sections received little attention. Comprehensive risk analysis has not been completely investigated To fill this gap, we will comprehensively compare four types of tra ffi c accidents between associated factors and find the contribution of di ff erent risk factors to the four di ff erent types of tra ffi c accidents at the weaving section. We divided the weaving section into five zones, which would be helpful in finding hazardous segments for four types of tra ffi c accidents at the weaving section. Tra ffi c accidents can be a ff ected by numerous potential factors, such as drivers’ behavior, weather condition, time, vehicle speed and so on. These factors include continuous and classified variables, which make correlation analysis between risk factors and tra ffi c accidents as a Multivariate Regression (MR) problem with multiple categorical variables. In addition, multinomial logistic regression is a statistical modeling technique on the premise that the probability for a dependent variable is related to a series of potential predictor variables [ 19 ], which is widely used to identify factors and predict the likelihood of an outcome [ 20 ]. For better understanding of the impact of risk factors on di ff erent types of tra ffi c accidents, this research extends the approach from two aspects, namely, (i) identifying six categories of risk factors such as drivers’ attributes, weather conditions, tra ffi c characteristics, driving behavior, vehicle types, and temporal-spatial distribution, (ii) applying the multinomial logistic regression approach to compare four types of tra ffi c accidents between associated factors This research makes the following contributions. Firstly, we precisely and comprehensively identify risk factors a ff ecting tra ffi c accidents at the weaving sections. Secondly, we propose a framework to establish the correlation between risk factors and tra ffi c accidents, and compare the impacts of various risk factors on di ff erent types of tra ffi c accidents The remainder of this paper is organized as follows. Section 2 reviews risk factors of tra ffi c accidents and the application of multinomial logistic regression. Section 3 chooses an observed weaving section, defines five analyzing zones, describes data collection, identifies six categories of risk factors and proposes the multinomial logistic regression as a method used in this research. Section 4 presents

Int. J. Environ. Res. Public Health 2019 , 16 , 1542 3 of 17 the calculation results obtained from the multinomial logistic regression approach. Section 5 analyzes and discusses the results, and the conclusions are drawn in Section 6 . 2. Literature Review 2.1. Risk Factor Analysis of Tra ffi c Accident at Weaving Sections Because of their serious consequences, tra ffi c safety issues have gained considerable attention from researchers [ 21 – 23 ]. As for tra ffi c accidents at the weaving section, important risk factors have been identified. For example, Cirillo analyzed the correlation between accident rates and configuration of weaving section based on the accident data collected in 1961, and the results showed that increasing the length of weaving areas, acceleration lanes, and deceleration lanes can reduce accident rates [ 24 ]. Fazio et al. carried out a safety analysis on Interstate 294 in the Chicago metropolitan area, which indicated that lane changing conflicts and following conflicts have a great e ff ect on crash rates at weaving sections [ 25 ]. Pulugurtha and Bhatt analyzed the influence of tra ffi c characteristics on crashes using the data of 581 crashes at 25 weaving sections in the Las Vegas metropolitan in 2000. The findings show that crash rates were low with weaving volume less than 15,000 vehicles per day, but high with weaving volume more than 50,000 vehicles per day, and increasing entry volume was the main factor causing the rise of improper lane change [ 1 ]. Penmetsa and Pulugurtha found that road features and drivers’ gender were the most two significant factors a ff ecting crash injury severity and the crash frequency at the weaving section using the crash data from 2011 to 2013 obtained from the Highway Safety Information System (HSIS) for the state of North Carolina [ 26 ]. Liu et al. studied the safety impacts of lane arrangements at the three types of weaving sections using generalized linear models [ 27 ]. Besides the above factors, weather conditions [ 28 , 29 ], vehicle types [ 30 , 31 ] and driving behavior [ 32 – 35 ] are also considered to be associated with tra ffi c crash risks. In addition, some other researchers studied tra ffi c crash prediction by establishing models to estimate crash likelihood at the weaving section. For instance, Wang et al. utilized a multilevel Bayesian logistic regression model to study crash likelihood using real-time crash data such as crash, geometric, and weather data at the weaving section, which indicated that the distance at which weaving turbulence no longer has impact had the highest risk of the crash [ 36 ]. To predict the frequency of accident occurrence, Caliendo et al employed Poisson distribution and negative multinomial regression models to establish relationships between tra ffi c crashes and tra ffi c flow, geometric infrastructure characteristics and environmental factors [ 37 ]. Kiattikomol et al. used a negative binomial regression modeling approach to develop separate models to predict numbers of crashes for di ff erent levels of crash severity for interchange segments and non-interchange segments respectively, based on the data obtained from North Carolina, USA [ 38 ]. From the previous studies, it is not di ffi cult to conclude that the number of tra ffi c accidents at weaving sections and their injury severity can be a ff ected by di ff erent risk factors. However, di ff erent groups of risk factors causing tra ffi c crashes were analyzed independently, and there were relatively few studies concentrating on the identification of potential factors associated with classified tra ffi c accidents. Moreover, the correlation between tra ffi c accident types and the accident location at weaving sections was rarely discussed. In view of this, it is necessary to establish a comprehensive risk factors system and the correlation between di ff erent types of tra ffi c accidents and their locations at the weaving section should be analyzed 2.2. Application of Multinomial Logistic Regression MR techniques refer to statistical methods that establish linear or nonlinear quantitative correlations among variables, which are used to discuss the dependence between an outcome and a ff ecting factors [ 39 , 40 ]. An increasing number of studies have focused on MR techniques, such as nonlinear regression [ 41 ], linear regression [ 42 ], stepwise regression [ 43 ], ridge regression [ 44 ], lasso regression [ 45 ], logistic regression [ 46 ] and so on. These methods can e ff ectively solve MR problems from di ff erent

Int. J. Environ. Res. Public Health 2019 , 16 , 1542 4 of 17 perspectives. As one of the most applicable logistic regression techniques, multinomial logistic regression can handle the case where the outcome variable is nominal with more than two levels. It is usually utilized to solve MR problems in various aspects, including medicine, economics, engineering and sociology, etc. For example, Kurt et al. studied risk factors a ff ecting coronary artery disease using multinomial logistic regression, which indicated that obesity, smoking status and age were the three most important factors causing coronary artery disease [ 47 ]. Lu et al. employed a multinomial logistic model depending upon the estimation of cumulative probability to identify the factors leading to the severity of tra ffi c accidents at Shanghai river-crossing tunnel, and the regression results showed that speed limit and driver’s gender were the two most important factors [ 48 ]. To predict company failures, Jabeur applied multinomial logistic regression to find the relationship between bankruptcy and 33 factors for two samples of healthy and failing companies [ 49 ]. Despite the wide range of applications of multinomial logistic regression, it is rare in literature to analyze risk factors a ff ecting di ff erent types of tra ffi c accidents at the weaving section. This research chooses multinomial logistic regression mainly considering two following advantages: (i) it does not need any assumption about the distribution of variables [ 50 ] and (ii) it is suitable for both continuous and categorical variables [ 51 ]. Furthermore, because risk factors (independent variables) were classified into six categories and tra ffi c accident types (dependent variables) were divided into four groups (rear-end, side wipe, collision with fixtures and rollover), multinomial logistic regression is a suitable method to solve the regression problem with multiple risk factors and multiple tra ffi c accident types in this research 3. Materials and Methods 3.1. Observation of a Weaving Section In this research, we make a field observation of a weaving section in Xi’an, a city of western China The observed weaving section is on the 3 rd Ring Road between Tian Wang Interchange and Ba Qiao Interchange. The weaving section is 624 m long, which has a three lane mainline with an auxiliary lane and one lane on / o ff ramps at both ends. Each lane is 3.75 m wide. It is a typical Type A weaving section. Location and configuration of the observed weaving section are illustrated in Figure 2 . Int. J. Environ. Res. Public Health 2019 , 16 , x 4 of 16 Int. J. Environ. Res. Public Health 2019 , 16 , x; doi: www.mdpi.com/journal/ijerph multinomial logistic regression can handle the case where the outcome variable is nominal with more than two levels. It is usually utilized to solve MR problems in various aspects, including medicine, economics, engineering and sociology, etc. For example, Kurt et al. studied risk factors affecting coronary artery disease using multinomial logistic regression, which indicated that obesity, smoking status and age were the three most important factors causing coronary artery disease [47]. Lu et al. employed a multinomial logistic model depending upon the estimation of cumulative probability to identify the factors leading to the severity of traffic accidents at Shanghai river-crossing tunnel, and the regression results showed that speed limit and driver’s gender were the two most important factors [48]. To predict company failures, Jabeur applied multinomial logistic regression to find the relationship between bankruptcy and 33 factors for two samples of healthy and failing companies [49]. Despite the wide range of applications of multinomial logistic regression, it is rare in literature to analyze risk factors affecting different types of traffic accidents at the weaving section. This research chooses multinomial logistic regression mainly considering two following advantages: (i) it does not need any assumption about the distribution of variables [50] and (ii) it is suitable for both continuous and categorical variables [51]. Furthermore, because risk factors (independent variables) were classified into six categories and traffic accident types (dependent variables) were divided into four groups (rear-end, side wipe, collision with fixtures and rollover), multinomial logistic regression is a suitable method to solve the regression problem with multiple risk factors and multiple traffic accident types in this research. 3. Materials and Methods 3.1. Observation of a Weaving Section In this research, we make a field observation of a weaving section in Xi’an, a city of western China. The observed weaving section is on the 3 rd Ring Road between Tian Wang Interchange and Ba Qiao Interchange. The weaving section is 624 m long, which has a three lane mainline with an auxiliary lane and one lane on/off ramps at both ends. Each lane is 3.75 m wide. It is a typical Type A weaving section. Location and configuration of the observed weaving section are illustrated in Figure 2. ( a ) ( b ) Figure 2. The observed weaving section. ( a ) Location of the observed weaving section; ( b ) Configuration of the observed weaving section . 3.2. Zone Definition Zone definition of a weaving section is necessary to identify the location of every single accident and obtain the spatial distribution regularities of accidents. Since the concentration of lane changing activity at the weaving section was in the first 200 m, Al-Jameel divided the 200 m segment of the whole M 60-J 2 weaving section throughout the Greater Manchester area in the UK into 4 equal zones [52]. It is observed that up to 90% of lane changing activities took place in the first 520 m of the observed weaving section. Using Al-Jameel’s definition, we divided the weaving Figure 2. The observed weaving section. ( a ) Location of the observed weaving section; ( b ) Configuration of the observed weaving section 3.2. Zone Definition Zone definition of a weaving section is necessary to identify the location of every single accident and obtain the spatial distribution regularities of accidents. Since the concentration of lane changing activity at the weaving section was in the first 200 m, Al-Jameel divided the 200 m segment of the whole M 60-J 2 weaving section throughout the Greater Manchester area in the UK into 4 equal zones [ 52 ]. It is observed that up to 90% of lane changing activities took place in the first 520 m of the observed weaving

Int. J. Environ. Res. Public Health 2019 , 16 , 1542 5 of 17 section. Using Al-Jameel’s definition, we divided the weaving section into five zones including four equal zones and an extra shorter zone in this research, plotted in Figure 3 . Int. J. Environ. Res. Public Health 2019 , 16 , x 5 of 16 Int. J. Environ. Res. Public Health 2019 , 16 , x; doi: www.mdpi.com/journal/ijerph section into five zones including four equal zones and an extra shorter zone in this research, plotted in Figure 3. Zone 1 (Z 1) is the first 130 m from entry point; Zone 2 (Z 2) is from 130 m to 260 m; Zone 3 (Z 3) is from 260 m to 390 m; Zone 4 (Z 4) is from 390 m to 520 m; Zone 5 (Z 5) is the remainder of the weaving section. Figure 3. Division of the observed weaving section into five zones 3.3. Data Collection Accident data and traffic data are both needed to identify risk factors of traffic accidents. Police accident records are usually adopted as reliable and important sources of traffic accident data [31]. In this research, traffic accident data was extracted from the Traffic Accident Database managed by Xi’an Public Security Bureau. We obtained 768 accident samples of the observed weaving section from 2016 to 2018. Each accident sample includes traffic accident type, drivers’ personal information, vehicle types, accident location features, weather conditions and the time of accidents. The samples included four types of traffic accidents: rear-end, side wipe, collision with fixtures and rollover. However, police accident records do not contain traffic characteristics when the accident happens. Hence, we employed video surveillance and tachometers to capture traffic data, including traffic volume, vehicle speed, traffic density and weaving ratio. Real-time traffic data was recorded 24 h every day during the three years and was transferred back to the laboratory. Traffic volume and vehicle speed can be collected directly, and traffic density was calculated using Equation (1). K = N L (1) where K is traffic density (vehicle/km); N is the number of vehicles (vehicle); L is the length of the lane (km). Weaving ratio is the percentage of the weaving vehicles out of the total number of the inflow vehicles to the section, which can be calculated using Equation (2). V R = Q W 1 + Q W 2 Q (2) where V R is weaving ratio (%); Q W 1 is ramp-to-mainline traffic volume (vehicle); Q W 2 is mainline-to-ramp traffic volume (vehicle); Q is the total traffic volume in the weaving section (vehicle). 3.4. Risk Factors Appropriate identification of risk factors affecting traffic accidents is necessary. From the existing research, driver attributes [53], weather conditions [54], traffic characteristics [55], driving behavior [56], temporal-spatial distribution [57] and vehicle types [58] can all influence the possibility of a traffic accident. We established the risk factor system based on the literature plus Figure 3. Division of the observed weaving section into five zones Zone 1 (Z 1) is the first 130 m from entry point; Zone 2 (Z 2) is from 130 m to 260 m; Zone 3 (Z 3) is from 260 m to 390 m; Zone 4 (Z 4) is from 390 m to 520 m; Zone 5 (Z 5) is the remainder of the weaving section 3.3. Data Collection Accident data and tra ffi c data are both needed to identify risk factors of tra ffi c accidents. Police accident records are usually adopted as reliable and important sources of tra ffi c accident data [ 31 ]. In this research, tra ffi c accident data was extracted from the Tra ffi c Accident Database managed by Xi’an Public Security Bureau. We obtained 768 accident samples of the observed weaving section from 2016 to 2018. Each accident sample includes tra ffi c accident type, drivers’ personal information, vehicle types, accident location features, weather conditions and the time of accidents. The samples included four types of tra ffi c accidents: rear-end, side wipe, collision with fixtures and rollover However, police accident records do not contain tra ffi c characteristics when the accident happens Hence, we employed video surveillance and tachometers to capture tra ffi c data, including tra ffi c volume, vehicle speed, tra ffi c density and weaving ratio. Real-time tra ffi c data was recorded 24 h every day during the three years and was transferred back to the laboratory Tra ffi c volume and vehicle speed can be collected directly, and tra ffi c density was calculated using Equation (1) K = N L (1) where K is tra ffi c density (vehicle / km); N is the number of vehicles (vehicle); L is the length of the lane (km) Weaving ratio is the percentage of the weaving vehicles out of the total number of the inflow vehicles to the section, which can be calculated using Equation (2) V R = Q W 1 + Q W 2 Q (2) where V R is weaving ratio (%); Q W 1 is ramp-to-mainline tra ffi c volume (vehicle); Q W 2 is mainline-to-ramp tra ffi c volume (vehicle); Q is the total tra ffi c volume in the weaving section (vehicle) 3.4. Risk Factors Appropriate identification of risk factors a ff ecting tra ffi c accidents is necessary. From the existing research, driver attributes [ 53 ], weather conditions [ 54 ], tra ffi c characteristics [ 55 ], driving behavior [ 56 ], temporal-spatial distribution [ 57 ] and vehicle types [ 58 ] can all influence the possibility of a tra ffi c

Int. J. Environ. Res. Public Health 2019 , 16 , 1542 6 of 17 accident. We established the risk factor system based on the literature plus two newly added risk factors, i.e., weaving ratio and accident location. The risk factor system consists of six categories including 16 risk factors as follows 3.4.1. Drivers’ Attributes Drivers’ age and gender are always considered as important factors in the research of tra ffi c accidents. According to di ff erent driving behaviors and skills, drivers’ age is classified into four groups: ≤ 25, 26–44, 45–64 and ≥ 65. Additionally, male and female drivers also have a di ff erent driving preference 3.4.2. Weather Conditions Weather condition is an external contributor to tra ffi c accidents, especially some extreme climatic conditions. From the data collection, eight di ff erent weather conditions were observed, i.e., sunny, cloudy, rainy, snowy, foggy, windy, dusty, hail [ 54 ]. 3.4.3. Tra ffi c Characteristics We used two main tra ffi c flow parameters, i.e., tra ffi c density and weaving ration to represent tra ffi c characteristics in the weaving section. Tra ffi c density is grouped into four categories: ≤ 10 vehicle / 100 m, 11–20 vehicle / 100 m, 21–30 vehicle / 100 m and ≥ 31 vehicle / 100 m. Weaving ration is also divided into four groups: ≤ 10%, 11–25%, 26–40% and ≥ 41% 3.4.4. Driving Behavior Speed is an important risk causing tra ffi c accidents, which is classified as ≤ 50 km / h, 51–80 km / h, 81–100 km / h and ≥ 101 km / h according to their di ff erent possibilities of causing an accident. At a weaving section, lane change is a common driving behavior in ramp-to-mainline and mainline-to-ramp tra ffi c flows, which easily leads to side wipe tra ffi c accidents 3.4.5. Vehicle Types From the 768 accident samples of the observed weaving section, five types of vehicles were involved in tra ffi c accidents, i.e., private car, minibus, bus, taxi and truck 3.4.6. Temporal-Spatial Distribution Seasons are grouped into four periods: spring, summer, autumn and winter. Time is divided into five categories: 00:00–06:59 a.m., 07:00–08:59 a.m., 09:00 a.m.–16:59 p.m., 17:00–19:59 p.m. and 20:00–23:59 p.m. 07:00–08:59 a.m. is the morning rush hour, and 17:00–19:59 p.m. is the evening rush hour. Weekends and weekdays are also classified. According to the zone definition of the weaving section, accident location is divided into five zones: Zone 1, Zone 2, Zone 3, Zone 4, and Zone 5 3.5. Methods Considering the advantages presented in Section 2.2 , the multinomial logistic regression is employed in this research, which is formulated in detail as follows It denotes that X i is the variable of tra ffi c accident a ff ecting factor i , i = 1, 2, . . , m , P j is the probability of tra ffi c accident type j , j = 0, 1, · · · , n − 1. j = 0 is used as the referent type of tra ffi c accident The regression relationship between X i and P j is formulated as P j = exp ( α j + P m i = 1 β i j · X i 1 + exp ( α j + P m i = 1 β i j · X i , j = 1, 2, · · · , n (3)

Int. J. Environ. Res. Public Health 2019 , 16 , 1542 7 of 17 where m is the number of risk factors; n is the number of tra ffi c accident types P j must be constrained as: n − 1 X j = 1 P j = 1 (4) Denote Y j = α j + β 1 j · X i + β 2 j · X i + · · · + β m j · X i = α j + P m i = 1 β i j · X i Where Y j is the total discrimination value which reflects the quantitative characteristics of i th tra ffi c accident a ff ecting factor; β i j is the coe ffi cient which reflects the degree of relevant independent variables X i ; α j is a constant Then, Equation (3) can be rewritten as: P j = exp Y j 1 + exp Y j (5) From Equation (5), exp Y j can be obtained as: exp Y j = P j 1 − P j (6) The natural logarithm is taken on Equation (6), and Y j can be formulated as: Y j = ln P j 1 − P j = α j + m X i = 1 β i j · X i (7) We use odd ration (OR) to estimate the e ff ects of di ff erent tra ffi c accident a ff ecting factors on the possibility of tra ffi c accident types. OR can be computed by Equation (8) OR i j = EXP ( β i j (8) where the odd ration of i -th tra ffi c accident a ff ecting factor to tra ffi c accident type j , which indicates the relative amount by which the odds of tra ffi c accidents increases (OR > 1) or decreases (OR < 1) when the value of a ff ecting factors increase per unit For the above multinomial logistic regression model, the significance of variables should be assessed. We employed the Wald test to achieve the significance testing, which is defined as β / SE (standard error) 4. Results The multinomial logistic regression model was applied using the collected data described in Appendix A , Table A 1 , which shows the number of four types of tra ffi c accidents associated with every single factor. Table 1 displays the multinomial logistic regression results, which indicates that the a ff ecting factors have di ff erent e ff ects on the four tra ffi c accident types. The detailed description of results is as follows.

Int. J. Environ. Res. Public Health 2019 , 16 , 1542 8 of 17 Table 1. Multinomial logistic regression results Factors Rear-End Side Wipe Collision with Fixtures Rollover β Wald OR β Wald OR β Wald OR β Wald OR Gender (Reference: Female) Male − 6.229 4.124 1.821 − 8.036 1.543 1.732 − 6.861 2.653 0.811 − 7.231 2.473 2.013 Age (Reference: ≤ 25) 26–44 − 9.137 11.763 1.123 − 10.51 2.854 1.265 − 9.618 6.685 1.521 − 9.899 6.065 2.223 45–64 − 4.215 6.432 0.743 − 4.814 0.894 0.548 − 4.425 3.275 0.682 − 4.547 2.891 0.873 ≥ 65 − 9.879 5.447 1.113 − 11.43 5.834 1.427 − 10.42 5.668 1.652 − 10.74 5.695 0.728 Weather Conditions (Reference: Sunny) Cloudy − 13.98 16.543 1.004 − 15.26 5.983 1.112 − 14.42 10.52 1.071 − 14.69 9.789 1.239 Rainy 13.74 8.762 2.687 11.48 6.936 3.431 12.95 7.721 1.652 12.49 7.594 4.871 Snowy − 17.13 2.871 5.432 − 22.05 4.991 2.721 − 18.85 4.079 6.137 − 19.86 4.227 4.652 Foggy − 9.725 1.762 4.247 − 13.53 3.746 2.689 − 11.06 2.893 2.148 − 11.84 3.031 2.651 Windy − 13.25 1.983 2.324 − 14.85 9.431 1.436 − 13.81 6.228 1.872 − 14.14 6.747 1.625 Dusty − 10.13 4.672 1.872 − 11.92 5.983 1.673 − 10.76 5.419 1.562 − 11.12 5.513 1.238 Hail − 11.01 6.432 1.004 − 12.27 6.783 1.121 − 11.45 6.632 1.105 − 11.71 6.656 1.217 Tra ffi c Density (Reference: ≤ 10 vehicle / 100 m) 11–20 vehicle / 100 m 10.75 11.325 2.315 8.461 4.093 2.422 9.951 7.203 0.621 9.482 6.699 1.012 21–30 vehicle / 100 m − 6.145 4.761 4.332 − 9.84 1.119 3.621 − 7.438 2.685 0.341 − 8.194 2.432 0.535 ≥ 31 vehicle / 100 m − 6.533 2.431 6.321 − 13.73 1.329 7.123 − 9.053 1.803 0.214 − 10.53 1.726 0.322 Weaving Ratio (Reference: ≤ 10%) 11–25% − 5.636 2.984 1.013 − 7.358 3.095 1.654 − 6.239 3.047 1.012 − 6.591 3.055 0.873 26–40% 3.758 4.872 1.103 0.26 0.678 3.543 2.534 2.481 1.033 1.818 2.189 2.451 ≥ 41% 9.72 2.743 1.024 2.325 7.774 7.512 7.132 5.611 1.154 5.62 5.961 0.231 Vehicle Speed (Reference: ≤ 50 km / h) 51–80 km / h − 7.986 1.834 1.533 − 9.394 4.875 2.312 − 8.479 3.567 1.276 − 8.767 3.779 1.467 81–100 km / h − 7.08 1.336 3.121 − 8.808 1.076 1.643 − 7.685 1.188 3.567 − 8.038 1.175 3.543 ≥ 101 km / h − 15.87 2.991 5.643 − 18.07 4.095 1.012 − 16.64 3.626 6.543 − 17.09 3.697 7.113 Lane Change (Reference: No Lane Change) Changing lanes − 15.08 5.982 2.143 − 24.39 3.775 7.124 − 18.34 4.724 6.641 − 20.24 4.571 7.214 Private Car (Reference: No private car involved) Private car involved − 16.26 11.437 4.717 − 17.89 9.453 5.435 − 16.84 10.31 3.612 − 17.17 10.17 5.498 Minibus (Reference: No minibus car involved) Minibus car involved − 4.989 0.763 0.887 − 6.961 1.734 1.912 − 5.679 1.316 1.654 − 6.082 1.384 1.784 Bus (Reference: No bus involved) Bus involved 2.876 1.218 0.723 1.7 0.673 1.211 2.464 0.907 1.109 2.224 0.869 1.045 Taxi (Reference: No taxi involved) Taxi involved 3.214 0.746 1.234 2.139 0.943 1.114 2.838 0.858 1.092 2.618 0.872 1.218 Truck (Reference: No truck involved) Truck involved − 8.944 2.198 4.019 − 11.17 4.874 0.617 − 9.722 3.723 2.215 − 10.18 3.909 0.485

Int. J. Environ. Res. Public Health 2019 , 16 , 1542 9 of 17 Table 1. Cont Factors Rear-End Side Wipe Collision with Fixtures Rollover β Wald OR β Wald OR β Wald OR β Wald OR Seasons (Reference: Spring) Summer − 6.236 1.043 1.012 − 7.346 4.098 2.035 − 6.625 2.784 1.108 − 6.852 2.997 1.715 Autumn − 6.345 4.15 1.123 − 7.608 3.004 1.187 − 6.787 3.497 1.165 − 7.045 3.417 0.898 Winter − 8.667 6.437 2.102 − 10.82 2.945 1.051 − 9.421 4.447 1.986 − 9.862 4.204 2.016 Time (Reference: 00:00-06:59) 07:00–08:59 − 11.16 4.983 2.451 − 14.06 5.843 1.762 − 12.18 5.473 1.263 − 12.77 5.533 0.832 09:00–16:59 − 10.37 2.336 1.873 − 12.43 6.657 1.932 − 11.09 4.799 1.784 − 11.51 5.104 1.073 17:00–19:59 5.613 4.776 1.763 3.737 1.002 1.943 4.956 2.625 1.672 4.573 2.362 0.733 20:00–23:59 6.554 8.984 1.032 5.457 1.431 1.176 6.17 4.679 2.154 5.946 4.153 2.149 Day of Week (Reference: Weekends) Weekdays 10.64 2.119 4.732 6.258 2.843 4.512 9.108 2.532 3.872 8.211 2.582 4.034 Accident Location (Reference: Zone 1) Zone 2 − 17.82 1.054 1.121 − 26.46 1.564 8.435 − 20.84 1.345 4.02 − 22.61 1.382 3.831 Zone 3 − 9.143 2.541 1.638 − 13.27 4.438 4.014 − 10.59 3.622 1.836 − 11.43 3.754 0.843 Zone 4 − 7.257 1.774 4.325 − 8.801 4.663 1.457 − 7.797 3.421 0.667 − 8.113 3.622 1.103 Zone 5 − 8.642 1.054 3.212 − 11.17 3.317 2.426 − 9.528 2.344 0.784 − 10.05 2.501 4.037 Note: Significant at 5% level.

Int. J. Environ. Res. Public Health 2019 , 16 , 1542 10 of 17 4.1. Rear-End Male drivers are 1.821 times more likely to involve in rear-end than female drivers. Compared to young drivers (age ≤ 25), drivers of age 26–44 and ≥ 65 have a higher risk of rear-end, but the age group of 45–64 has the lowest risk. Snow and fog are the two most significant weather conditions, which are more inclined to cause rear-end than other weather conditions. The higher the tra ffi c density is, the higher the risk it will have. It seems that weaving ratio does not have an obvious impact on rear-end, because the ORs show little variation with the change of weaving ratio. Tra ffi c density of ≥ 31 vehicle / 100 m is associated with the highest possibility of rear-end (OR = 6.321). In addition, lane change behavior is 2.143 times more probably to bring about rear-end than no lane change behavior. Private cars, taxis and trucks are more inclined to be involved in rear-end, especially trucks. Compared to other seasons, more rear-end tra ffi c accidents occur in winter. As for time, morning rush hour (07:00–08:59) exhibits the highest risk for rear-end, while the time period 00:00–06:59 has the lowest risk. Rear-end is less likely to occur during weekends. It is found that Zone 4 and Zone 5 correlated with a higher chance of rear-end at the weaving section 4.2. Side Wipe Male drivers are more inclined to cause a side wipe than female drivers. Compared to young drivers (age ≤ 25), drivers of age ≥ 65 are 1.427 times more likely to be involved in a side wipe, but the other two age groups have lower risks. Rain is found to be the most likely condition to lead to a side wipe compared to the other weather conditions. There is a positive correlation between tra ffi c density and side wipe risk. The higher the weaving ratio is, the higher possibility the side wipe will have. Weaving ration of ≥ 41% is the most likely to bring about a side wipe (OR = 7.512). Furthermore, compared to low speed, higher speed has higher chances of a side wipe, but middle speed 51–80 km / h has the highest risk. Lane change is also an important factor a ff ecting the side wipe. Private cars and taxis are more associated with side wipes, while trucks are less likely to be involved in the side wipe. Side wipe is more likely to happen in summer and in the evening rush hour (17:00–19:59), but less likely at weekends. There is a higher risk of side wipe in Zone 2 and Zone 3 at the weaving section 4.3. Collision with Fixtures Interestingly, unlike the other tra ffi c accident types, female drivers have a higher possibility of collision with fixtures than male drivers. Drivers of age 26–44 and ≥ 65 have a higher risk of collision with fixtures, but drivers of age 45–64 have a lower risk compared to young drivers (age ≤ 25). Snowy weather has the highest possibility to bring about the collision with fixtures (OR = 6.137). Higher tra ffi c density will reduce the probability of collision with fixtures. Risk of collision with fixtures becomes higher as the weaving ratio increases, but the ORs show little variation. It is obtained that when speed is more than 101 km / h, vehicles have the highest risk of being associated with the collision with fixtures compared to other vehicle speed categories. Lane change behavior is also an important factor causing the collision with fixtures. Private cars are more inclined to be involved in a collision with fixtures. A collision with fixtures is more likely to occur in winter and in the time period 09:00–16:59, but less likely at weekends. Zone 2 and Zone 5 of the weaving section are associated with a higher chance of collision with fixtures than the other three zones 4.4. Rollover Male drivers are 2.013 times more likely to be involved in the rollover. Drivers of age 26–44 are more inclined to cause a rollover, but drivers of age 45–64 and ≥ 65 have lower risks compared to young drivers (age ≤ 25). Rainy weather has the greatest impact on rollover (OR = 4.871). Higher tra ffi c density shows a lower probability of rollover. Middle weaving ratio (26–40%) has a higher risk of rollover, while lower weaving ratio (11–25%) and higher weaving ratio ( ≥ 41%) are correlated with lower risk. Risk of rollovers increases with vehicle speed. Lane change indicates a much higher risk of

Int. J. Environ. Res. Public Health 2019 , 16 , 1542 11 of 17 rollover than no lane change. Private cars are more inclined to be involved in the rollover, but trucks present a lower risk of rollover. It is found that rollover is more likely to occur in summer, especially in winter. The time period of 20:00–23:59 has the highest risk of rollover, while the time period of 17:00–19:59 shows the lowest risk. In addition, rollover is more likely to happen during weekdays than weekends. There is a higher risk of rollover in Zone 2 and Zone 5 at the weaving section 5. Discussion 5.1. Drivers’ Attributes Drivers’ gender is found to be an important factor a ff ecting tra ffi c accidents at the weaving section [ 59 ]. From the statistical analysis, we know that male drivers have a higher possibility of involvement in rear-end, side wipe and rollover than female drivers, but for collision with fixtures, female drivers have higher risks than male drivers According to the results, drivers aged 26–44 and ≥ 65 are more inclined to be involved in rear-end, side wipe and rollover at the weaving section. This is probably because drivers aged 26–44 usually have bad driving behaviors such as speeding, using cell phones while driving, frequent lane change, etc. and drivers aged ≥ 65 have a decline of driving skills Regarding this concern, the criteria for driving license issuance for older drivers requires further analysis and stricter measures should be adopted to prevent drink driving, distracted driving, road rage, etc 5.2. Weather Conditions Bad weather conditions are always identified as a high risk, which may lead to a severe tra ffi c accident injury. Compared to sunny days, other weather conditions have higher risks to a di ff erent extent at the weaving section. Snow has the highest possibility of rear-end and collision with fixtures, which is probably due to the slippery pavement. While rain has the highest risk of side wipe and rollover, because rain has a great impact on drivers’ vision and sight As a response, drivers are advised to drive less and not to drive faster than the speed limit during bad weather conditions. Additionally, drivers should check their tires regularly to ensure the grip of the tires 5.3. Tra ffi c Characteristics According to the results, higher tra ffi c density is more associated with rear-end and side wipe, because higher tra ffi c density means a shorter average space headway and more congested tra ffi c flow There is an inverse relationship between tra ffi c density and the risk of collision with fixtures, as well as rollover. Weaving ratio is an important parameter to describe the characteristics of tra ffi c flow at the weaving section, which has a great risk of side wipe and rollover, but has little e ff ect on rear-end and collision with fixtures Accordingly, it is advisable for drivers to keep a safe distance from vehicles in front of them, and prepare for a lane change ahead of time when they enter the ramp from mainline and vice versa. The aggressive competition for a lane-changing opportunity is forbidden at the weaving section 5.4. Driving Behavior Many existing research results show that speeding and frequent lane change are the two most common dangerous driving behaviors, which cause most of the tra ffi c accidents. From the results, it is revealed that higher vehicle speed has a higher chance of causing rear-end, collision with fixtures and rollover, but has a lower risk of side wipe. Lane change has a great e ff ect on side wipe and rollover at the weaving section. Hence, it is necessary to promote mandated speed limits and no overtaking at the weaving section and to build more than one auxiliary lane for the weaving section if possible.

Int. J. Environ. Res. Public Health 2019 , 16 , 1542 12 of 17 5.5. Vehicle Types Private cars account for 87% of the total tra ffi c volume at the observed weaving section, which has the highest possibility of involvement in tra ffi c accidents, i.e., they are 4.717 times more likely to be involved in rear-end, 5.435 times in side wipe, 3.612 times in collision with fixtures and 5.498 times in rollover than non-private cars. Because trucks always have a longer braking distance, they are more associated with rear-end compared to the other three types of tra ffi c accidents at the weaving section. Since taxis usually have a high speed and frequent lane change, they are also often at risk of tra ffi c accidents. In view of this, tra ffi c regulations could be considered to separate passenger cars and trucks. In addition, some heavy trucks may only be allowed to drive along specialized lanes or during a certain times at the weaving section 5.6. Temporal-Spatial Distribution Rear-end, collision with fixtures and rollover are more inclined to happen in winter, because bad weather conditions such as rain, snow, fog and hail often occur in winter, which are risks for tra ffi c accidents, while side wipe is more likely to happen in summer. Because tra ffi c density is usually high in rush hours, morning rush hour often has a higher risk of rear-end, but it is found that evening rush hour has a higher risk of side wipe. Collision with fixtures and rollover occur in the period 20:00–23:59 most often, which is probably due to drowsy driving. All four types of tra ffi c accidents are much more likely to happen during weekdays than weekends. Tra ffi c accidents have significant spatial distribution characteristics at the weaving section in this research, which is rarely studied in the previous literature. It is found that rear-end is more likely to occur in Zone 4 and Zone 5, side wipe has a higher possibility of happening in Zone 2 and Zone 3, collision with fixtures and rollover are more inclined to happen in Zone 2 and Zone 5 As a result, tra ffi c demand management policies such as flexible work time should be designed to shift tra ffi c from peak periods. Extension of the weaving length, improvement of geometric conditions, avoiding horizontal curve or vertical curve, etc. are also should be considered to enhance the safety level of the weaving section 6. Conclusions To solve the problem of land shortage, lots of interchanges have been built in China’s metropolitan areas, which increased transport mobility greatly. As an important configuration of interchanges, weaving section is likely to be involved in more tra ffi c accidents, which may bring about severe casualties and significant economic losses. Analysis of risk factors has become necessary to prevent various types of tra ffi c accidents at weaving sections. In view of this, this research established a risk factor identification and analysis framework of tra ffi c accidents at weaving sections using multinomial logistic regression. Correlation between six categories of 16 risk factors (drivers’ attributes, weather conditions, tra ffi c characteristics, driving behavior, vehicle types and temporal-spatial distribution) and four types of tra ffi c accidents (rear-end, side wipe, collision with fixtures and rollover) was identified based on 768 accident samples of an observed weaving section in Xi’an, China from 2016 to 2018. Di ff erent significance of 16 risk factors in the four types of tra ffi c accidents was compared. The main results are concluded as follows: (1) Tra ffi c accidents at the weaving section are mainly a ff ected by factors such as drivers’ gender and age, weather condition, tra ffi c density, weaving ratio, vehicle speed, lane change behavior, private cars, season, time period, day of week and accident location, but these factors have di ff erent e ff ects on the four tra ffi c accident types (2) Tra ffi c density of ≥ 31 vehicle / 100 m has the highest risk of causing rear-end, weaving ration of ≥ 41% has the highest possibility of bringing about a side wipe, collision with fixtures is the most likely to happen in snowy weather, and rollover is the most likely to occur in rainy weather.

Int. J. Environ. Res. Public Health 2019 , 16 , 1542 13 of 17 Due to their serious results, tra ffi c accidents attract tremendous attention from researchers However, few studies have yet analyzed and compared the di ff erent significance of risk factors in di ff erent types of tra ffi c accidents at the weaving section. This research intends to provide a reference to improve the safety of weaving sections in China. However, there are two limitations in this research. Firstly, this research focuses on the tra ffi c accidents at the Type A weaving section, but does not consider other types of weaving sections (Types B and C), which have di ff erent configuration characteristics Secondly, the injury severity of tra ffi c accidents was not involved in this research. These limitations will be taken into account in the following studies Author Contributions: X.M. designed research goals and wrote the manuscript; C.Y. designed research methods; J.G. collected and analyzed the data and S.Z. edited the manuscript Funding: This research was funded by Social Science Research Program of Ministry of Education in China (Grant Number 16 XJCZH 002) and supported by the Fundamental Research Funds for the Central Universities (Grant Number 310823170657 and 300102238501) and National Natural Science Foundation of China (Grant Number 71701022) and Natural Science Basic Research Plan in Shaanxi Province of China (Grant Number 2018 JQ 7002) and National Key R & D project (Grant Number 2017 YFC 0803906) Conflicts of Interest: The authors declare no conflicts of interest Appendix A Table A 1. Descriptive statistics of variables (number of tra ffi c accidents) Risk Factors Variables Rear-End Side Wipe Collision with Fixtures Rollover Drivers’ Attributes Gender Male 131 156 38 24 Female 182 194 28 15 Age ≤ 25 38 52 12 8 26-44 138 143 24 12 45-64 119 128 27 14 ≥ 65 18 27 3 5 Weather Conditions Weather Conditions Sunny 49 56 6 4 Cloudy 41 48 10 6 Rainy 85 91 15 9 Snowy 44 49 12 8 Foggy 19 31 8 5 Windy 25 25 6 4 Dusty 22 22 5 2 Hail 28 28 4 1 Tra ffi c Characteristics Tra ffi c Density ≤ 10 vehicle / 100 m 39 43 26 18 11-20 vehicle / 100 m 64 71 24 16 21-30 vehicle / 100 m 115 127 10 2 ≥ 31 vehicle / 100 m 95 109 6 3 Weaving Ratio ≤ 10% 21 19 13 5 11%–25% 50 48 9 4 26%–40% 115 119 18 13 ≥ 41% 127 164 26 17 Driving Behavior Vehicle Speed ≤ 50 km / h 36 43 6 3 51-80 km / h 54 63 9 4 81-100 km / h 98 109 23 11 ≥ 101 km / h 125 135 26 21 Lane Change No lane change 149 44 23 8 Changing lanes 164 306 43 31

Int. J. Environ. Res. Public Health 2019 , 16 , 1542 14 of 17 Table A 1. Cont Risk Factors Variables Rear-End Side Wipe Collision with Fixtures Rollover Vehicle Types Private Car No private car involved 39 51 12 6 Private car involved 274 299 54 33 Minibus No minibus car involved 248 288 55 30 Minibus car involved 65 62 11 9 Bus No bus involved 287 312 53 28 Bus involved 26 38 13 11 Taxi No taxi involved 266 312 55 31 Taxi involved 47 38 11 8 Truck No truck involved 288 322 60 35 Truck involved 25 28 6 4 Temporal Spatial Distribution Seasons Spring 73 78 14 9 Summer 76 84 15 9 Autumn 71 80 14 8 Winter 93 108 23 11 Time 00:00-06:59 13 15 2 2 07:00-08:59 83 93 17 10 09:00-16:59 113 127 24 14 17:00-19:59 61 68 13 8 20:00-23:59 43 47 10 5 Day of Week Weekends 134 150 28 17 Weekdays 179 200 38 22 Accident Location Zone 1 8 30 12 4 Zone 2 70 138 17 10 Zone 3 63 118 17 9 Zone 4 94 54 8 4 Zone 5 78 10 12 12 References 1 Pulugurtha, S.S.; Bhatt, J. Evaluating the role of weaving section characteristics and tra ffi c on crashes in weaving areas Tra ffi c Inj. Prev 2010 , 11 , 104–113. [ CrossRef ] [ PubMed ] 2 Golob, T.F.; Recker, W.W.; Alvarez, V.M. Safety aspects of freeway weaving sections Transp. Res. Part A: Policy Pract 2004 , 38 , 35–51. [ CrossRef ] 3 Jin, W.-L. A kinematic wave theory of lane-changing tra ffi c flow Transp. Res. Part B: Methodol 2010 , 44 , 1001–1021. [ CrossRef ] 4 Hidas, P. Modelling vehicle interactions in microscopic simulation of merging and weaving Transp. Res Part C: Emerg. Technol 2005 , 13 , 37–62. [ CrossRef ] 5 Tanaka, S.; Hasegawa, N.; Iizuka, D.; Nakamura, F. Evaluation of vehicle control algorithm to avoid conflicts in weaving sections under fully-controlled condition in urban expressway Transp. Res. Proced 2017 , 21 , 199–207. [ CrossRef ] 6 Md Diah, J.; Abdul Rahman, M.; Adnan, M.A.; Hooi Ling, K. Modeling the relationship between geometric design and weaving section flow process of conventional roundabouts J. Transp. Eng 2011 , 137 , 980–986 [ CrossRef ] 7 Hav â rneanu, G.M.; Hav â rneanu, C.E. When norms turn perverse: Contextual irrationality vs. rational tra ffi c violations Transp. Res. Part F: Tra ffi c Psychol. Behav 2012 , 15 , 144–151. [ CrossRef ] 8 Rundmo, T.; Iversen, H. Risk perception and driving behaviour among adolescents in two Norwegian counties before and after a tra ffi c safety campaign Saf. Sci 2004 , 42 , 1–21. [ CrossRef ] 9 Plug, C.; Xia, J.C.; Caulfield, C. Spatial and temporal visualisation techniques for crash analysis Accid. Anal Prev 2011 , 43 , 1937–1946. [ CrossRef ] [ PubMed ] 10 Yeung, J.S.; Wong, Y.D. Road tra ffi c accidents in Singapore expressway tunnels Tunn. Undergr. Space Technol 2013 , 38 , 534–541. [ CrossRef ]

Int. J. Environ. Res. Public Health 2019 , 16 , 1542 15 of 17 11 Abdel-Aty, M.A.; Radwan, A.E. Modeling tra ffi c accident occurrence and involvement Accid. Anal. Prev 2000 , 32 , 633–642. [ CrossRef ] 12 Dinu, R.; Veeraragavan, A. Random parameter models for accident prediction on two-lane undivided highways in India J. Saf. Res 2011 , 42 , 39–42. [ CrossRef ] 13 Loo, B.P. The identification of hazardous road locations: A comparison of the blacksite and hot zone methodologies in Hong Kong Int. J. Sustain. Transp 2009 , 3 , 187–202. [ CrossRef ] 14 Mart í n, L.; Baena, L.; Garach, L.; L ó pez, G.; de Oña, J. Using data mining techniques to road safety improvement in Spanish roads Proced. Soc. Behav. Sci 2014 , 160 , 607–614. [ CrossRef ] 15 Minderhoud, M.M.; Bovy, P.H. Extended time-to-collision measures for road tra ffi c safety assessment Accid. Anal. Prev 2001 , 33 , 89–97. [ CrossRef ] 16 Pasquini, A.; Pozzi, S. Evaluation of air tra ffi c management procedures—Safety assessment in an experimental environment Reliab. Eng. Syst. Saf 2005 , 89 , 105–117. [ CrossRef ] 17 Hänninen, M. Bayesian networks for maritime tra ffi c accident prevention: Benefits and challenges Accid. Anal Prev 2014 , 73 , 305–312. [ CrossRef ] 18 Salvarani, C.P.; Colli, B.O.; J ú nior, C.G.C. Impact of a program for the prevention of tra ffi c accidents in a Southern Brazilian city: A model for implementation in a developing country Surg. Neurol 2009 , 72 , 6–13 [ CrossRef ] [ PubMed ] 19 Tu, J.V. Advantages and disadvantages of using artificial neural networks versus logistic regression for predicting medical outcomes J. Clin. Epidemiol 1996 , 49 , 1225–1231. [ CrossRef ] 20 Yokoyama, M.; Yanagisawa, M. Logistic regression analysis of multiple interosseous hand muscle activation using surface electromyography during isometric finger tasks J. Electromyogr. Kinesiol 2019 , 44 , 117–123 [ CrossRef ] 21 Brow, G.R. Tra ffi c conflicts for road user safety studies Can. J. Civ. Eng 1994 , 21 , 1–15. [ CrossRef ] 22 Ulleberg, P. Personality subtypes of young drivers. Relationship to risk-taking preferences, accident involvement, and response to a tra ffi c safety campaign Transp. Res. Part F: Tra ffi c Psychol. Behav 2001 , 4 , 279–297. [ CrossRef ] 23 Lee, C.; Hellinga, B.; Saccomanno, F. Evaluation of variable speed limits to improve tra ffi c safety Transp. Res Part C: Emerg. Technol 2006 , 14 , 213–228. [ CrossRef ] 24 Cirillo, J.A. The relationship of accidents to length of speed-change lanes and weaving areas on interstate highways Highw. Res. Rec 1970 , 312 , 17–32 25 Fazio, J.; Holden, J.; Rouphail, N.M. Use of freeway conflict rates as an alternative to crash rates in weaving section safety analyses Transp. Res. Rec 1993 , 1401 , 61–69 26 Penmetsa, P.; Pulugurtha, S.S. Modeling crash injury severity by road feature to improve safety Tra ffi c Inj Prev 2018 , 19 , 102–109. [ CrossRef ] [ PubMed ] 27 Liu, P.; Chen, H.; Lu, J.J.; Cao, B. How lane arrangements on freeway mainlines and ramps a ff ect safety of freeways with closely spaced entrance and exit ramps J. Transp. Eng 2009 , 136 , 614–622. [ CrossRef ] 28 Fridstrøm, L.; Ifver, J.; Ingebrigtsen, S.; Kulmala, R.; Thomsen, L.K. Measuring the contribution of randomness, exposure, weather, and daylight to the variation in road accident counts Accid. Anal. Prev 1995 , 27 , 1–20 [ CrossRef ] 29 Golob, T.F.; Recker, W.W. Relationships among urban freeway accidents, tra ffi c flow, weather, and lighting conditions J. Transp. Eng 2003 , 129 , 342–353. [ CrossRef ] 30 Chang, L.-Y.; Wang, H.-W. Analysis of tra ffi c injury severity: An application of non-parametric classification tree techniques Accid. Anal. Prev 2006 , 38 , 1019–1027. [ CrossRef ] 31 Zhang, G.; Yau, K.K.; Chen, G. Risk factors associated with tra ffi c violations and accident severity in China Accid. Anal. Prev 2013 , 59 , 18–25. [ CrossRef ] [ PubMed ] 32 Alonso, F.; Esteban, C.; Montoro, L.; Serge, A. Conceptualization of aggressive driving behaviors through a Perception of aggressive driving scale (PAD) Transp. Res. Part F: Tra ffi c Psychol. Behav 2019 , 60 , 415–426 [ CrossRef ] 33 Evans, L.; Wasielewski, P. Risky driving related to driver and vehicle characteristics Accid. Anal. Prev 1983 , 15 , 121–136. [ CrossRef ] 34 Fergusson, D.; Swain-Campbell, N.; Horwood, J. Risky driving behaviour in young people: Prevalence, personal characteristics and tra ffi c accidents Aust. N. Z. J. Public Health 2003 , 27 , 337–342. [ CrossRef ] [ PubMed ]

Int. J. Environ. Res. Public Health 2019 , 16 , 1542 16 of 17 35 Ahammed, M.A.; Hassan, Y.; Sayed, T.A. Modeling driver behavior and safety on freeway merging areas J. Transp. Eng 2008 , 134 , 370–377. [ CrossRef ] 36 Wang, L.; Abdel-Aty, M.; Shi, Q.; Park, J. Real-time crash prediction for expressway weaving segments Transp. Res. Part C: Emerg. Technol 2015 , 61 , 1–10. [ CrossRef ] 37 Caliendo, C.; Guida, M.; Parisi, A. A crash-prediction model for multilane roads Accid. Anal. Prev 2007 , 39 , 657–670. [ CrossRef ] 38 Kiattikomol, V.; Chatterjee, A.; Hummer, J.E.; Younger, M.S. Planning level regression models for prediction of crashes on interchange and noninterchange segments of urban freeways J. Transp. Eng 2008 , 134 , 111–117 [ CrossRef ] 39 Sauerbrei, W.; Meier-Hirmer, C.; Benner, A.; Royston, P. Multivariable regression model building by using fractional polynomials: Description of SAS, STATA and R programs Comput. Stat. Data Anal 2006 , 50 , 3464–3485. [ CrossRef ] 40 Atici, U. Prediction of the strength of mineral admixture concrete using multivariable regression analysis and an artificial neural network Expert Syst. Appl 2011 , 38 , 9609–9618. [ CrossRef ] 41 Tsekouras, G.; Dialynas, E.; Hatziargyriou, N.; Kavatza, S. A non-linear multivariable regression model for midterm energy forecasting of power systems Electr. Power Syst. Res 2007 , 77 , 1560–1568. [ CrossRef ] 42 Razi, M.A.; Athappilly, K. A comparative predictive analysis of neural networks (NNs), nonlinear regression and classification and regression tree (CART) models Expert Syst. Appl 2005 , 29 , 65–74. [ CrossRef ] 43 Liao, X.; Li, Q.; Yang, X.; Zhang, W.; Li, W. Multiobjective optimization for crash safety design of vehicles using stepwise regression model Struct. Multidiscip. Optim 2008 , 35 , 561–569. [ CrossRef ] 44 Muniz, G.; Kibria, B.G. On some ridge regression estimators: An empirical comparisons Commun. Stat Simul. Comput 2009 , 38 , 621–630. [ CrossRef ] 45 Hans, C. Model uncertainty and variable selection in Bayesian lasso regression Stat. Comput 2010 , 20 , 221–229. [ CrossRef ] 46 Hayes, A.F.; Matthes, J. Computational procedures for probing interactions in OLS and logistic regression: SPSS and SAS implementations Behav. Res. Methods 2009 , 41 , 924–936. [ CrossRef ] 47 Kurt, I.; Ture, M.; Kurum, A.T. Comparing performances of logistic regression, classification and regression tree, and neural networks for predicting coronary artery disease Expert Syst. Appl 2008 , 34 , 366–374 [ CrossRef ] 48 Lu, J.J.; Xing, Y.; Wang, C.; Cai, X. Risk factors a ff ecting the severity of tra ffi c accidents at Shanghai river-crossing tunnel Tra ffi c Inj. Prev 2016 , 17 , 176–180. [ CrossRef ] 49 Jabeur, S.B. Bankruptcy prediction using partial least squares logistic regression J. Retail. Consum. Serv 2017 , 36 , 197–202. [ CrossRef ] 50 Bham, G.H.; Javvadi, B.S.; Manepalli, U.R. Multinomial logistic regression model for single-vehicle and multivehicle collisions on urban US highways in Arkansas J. Transp. Eng 2011 , 138 , 786–797. [ CrossRef ] 51 Dreiseitl, S.; Ohno-Machado, L. Logistic regression and artificial neural network classification models: A methodology review J. Biomed. Inform 2002 , 35 , 352–359. [ CrossRef ] 52 Al-Jameel, H. Characteristics of the driver behaviour in weaving sections: Empirical study Int. Eng 2013 , 2 , 1430–1446 53 Toran Pour, A.; Moridpour, S.; Tay, R.; Rajabifard, A. Influence of pedestrian age and gender on spatial and temporal distribution of pedestrian crashes Tra ffi c Inj. Prev 2018 , 19 , 81–87. [ CrossRef ] 54 Zhao, J.; Deng, W. Tra ffi c accidents on expressways: New threat to China Tra ffi c Inj. Prev 2012 , 13 , 230–238 [ CrossRef ] 55 Bai, L.; Chan, C.-Y.; Liu, P.; Xu, C. Identifying factors a ff ecting the safety of mid-block bicycle lanes considering mixed 2-wheeled tra ffi c flow Tra ffi c Inj. Prev 2017 , 18 , 761–766. [ CrossRef ] 56 Wang, Y.; Bao, S.; Du, W.; Ye, Z.; Sayer, J.R. A spectral power analysis of driving behavior changes during the transition from nondistraction to distraction Tra ffi c Inj. Prev 2017 , 18 , 826–831. [ CrossRef ] 57 Liu, C.; Sharma, A. Exploring spatio-temporal e ff ects in tra ffi c crash trend analysis Anal. Methods Acid. Res 2017 , 16 , 104–116. [ CrossRef ]

Int. J. Environ. Res. Public Health 2019 , 16 , 1542 17 of 17 58 Milton, J.C.; Shankar, V.N.; Mannering, F.L. Highway accident severities and the mixed logit model: An exploratory empirical analysis Accid. Anal. Prev 2008 , 40 , 260–266. [ CrossRef ] [ PubMed ] 59 Kim, K.; Brunner, I.M.; Yamashita, E. Modeling fault among accident—Involved pedestrians and motorists in Hawaii Accid. Anal. Prev 2008 , 40 , 2043–2049. [ CrossRef ] © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http: // creativecommons.org / licenses / by / 4.0 / ).