An accident severity analysis method and system based on mixed linear model

Abstract

The application belongs to the technical field of traffic accidents, and provides an accident severity analysis method and system based on a mixed linear model. The method comprises the following steps: acquiring personnel data, vehicle data, road data and environment data, and extracting personnel features, vehicle features, road features and environment features; using a general linear model to analyze whether the personnel features, vehicle features, road features and environment features have a significant influence on the severity of traffic accidents, and obtaining hierarchical clustering factors with a significance less than a set threshold; using the mixed linear model to analyze the influence of the in-vehicle personnel features on the severity of traffic accidents by taking the personnel features as fixed effects and the hierarchical clustering factors as random effects, so as to reduce the clustering of the hierarchical clustering factors. The application reduces the influence of factors such as vehicles, roads and environments on the severity of traffic accidents by comprehensively considering the correlation between the influencing factors and the accident risk.

Description

Technical Field

[0001] This invention belongs to the field of traffic accident technology, and in particular relates to a method and system for analyzing the severity of accidents based on a hybrid linear model. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] Domestic and international studies have confirmed that traffic accidents are influenced by multiple factors related to personnel characteristics, such as driver age, gender, fatigue level, number of passengers, and passenger age, gender, and seat position. Keli A. Braitman et al., based on the quasi-induced exposure method and logistic regression analysis, established functions of driver age and gender, as well as passenger age and gender, to predict the probability of fatal accidents. Aimee E. Cox et al. compared the accident risk trends of drivers aged 70 and above with those aged 35-54 from 1997 to 2018, finding a significant increase in fatal accidents among middle-aged drivers, while the rate decreased or remained stable among older drivers. Xu et al. used a stochastic parametric model to estimate the complex association between driver intentional or unintentional behavior and the severity of collision injuries. Li Ruirui et al. used a two-level logit model to comprehensively analyze the impact of fleet characteristics and various driver personal characteristics on bus accident risk, and ranked significant variables by importance based on the average accuracy reduction method of random forests. Haoqiang Fu et al. conducted a stepwise analysis of driver and passenger characteristics, studying the impact of passenger age and gender on the fatal accident risk of young drivers. Leon... Villavicencio et al. conducted a cross-sectional analysis of police-reported car accidents in the United States from 2016 to 2019, studying the impact of passenger number and age on the risk of death for adolescent drivers; Zhan Junjun used collision morphology, road type, and collision object as influencing variables, and analyzed the degree of injury of occupants in various seat positions in road traffic accidents involving passenger cars according to injury grading standards and multivariate analysis of variance; Nathaniel C Briggs et al. compared seat belt usage in driver and passenger seat positions, finding that high school students aged 16 and above in the United States were significantly less likely to wear seat belts as passengers than as drivers; Yang Ye selected independent variables from basic accident information, accident personnel data, and on-site environmental data, and used a binomial logit model to establish models of the severity of injury for drivers and passengers respectively to analyze and compare the injury situation of the two; Chris Lee et al.'s analysis using a bivariate probability model showed a strong correlation between passengers and collision characteristics. Zhang Qianyi et al. used chi-square test and binary logistic regression equation to comprehensively study the impact of motorcycles carrying passengers on driving behavior and traffic safety from four aspects: passenger carrying tendency, accident liability determination, accident type and accident severity. Peter Kiteywo Sisimwo et al. conducted a cross-sectional analysis of 341 patients with motorcycle collision injuries and found that wearing a helmet can effectively prevent head injuries to victims.

[0004] However, most existing studies only discuss the relationship between the characteristics of vehicle occupants and accident risk. Most studies use logistic regression analysis, which usually ignores the influence of other factors such as vehicles, roads, and environment on traffic accidents. Therefore, the research results may be biased. Summary of the Invention

[0005] To address the technical problems mentioned above, this invention provides a method and system for analyzing accident severity based on a hybrid linear model. This method comprehensively considers the relationship between various influencing factors and accident risks, reduces the impact of factors such as vehicles, roads, and the environment on the severity of traffic accidents, and clarifies the influence patterns of human factors. This provides theoretical support for safety education for drivers and passengers, timely restricts unsafe behaviors of drivers and passengers, and improves the overall safety situation.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] The first aspect of the present invention provides a method for accident severity analysis based on a hybrid linear model.

[0008] A method for accident severity analysis based on a hybrid linear model, comprising:

[0009] Acquire personnel data, vehicle data, road data, and environmental data, and extract personnel characteristics, vehicle characteristics, road characteristics, and environmental characteristics;

[0010] Using a general linear model, we analyzed whether personnel characteristics, vehicle characteristics, road characteristics, and environmental characteristics have a significant impact on the severity of traffic accidents, and obtained hierarchical clustering factors with significance less than a set threshold.

[0011] By treating occupant characteristics as a fixed effect and hierarchical clustering factors as random effects, a mixed linear model is used to analyze the impact of occupant characteristics on the severity of traffic accidents, thereby reducing the clustering of hierarchical clustering factors.

[0012] A second aspect of the present invention provides an accident severity analysis system based on a hybrid linear model.

[0013] An accident severity analysis system based on a hybrid linear model includes:

[0014] The data acquisition module is configured to acquire personnel data, vehicle data, road data, and environmental data, and extract personnel characteristics, vehicle characteristics, road characteristics, and environmental characteristics.

[0015] The general linear model processing module is configured to: use a general linear model to analyze whether personnel characteristics, vehicle characteristics, road characteristics, and environmental characteristics have a significant impact on the severity of traffic accidents, and obtain hierarchical clustering factors with significance less than a set threshold.

[0016] The mixed linear model processing module is configured to treat personnel characteristics as fixed effects and hierarchical clustering factors as random effects, and use a mixed linear model to analyze the impact of in-vehicle personnel characteristics on the severity of traffic accidents, thereby reducing the clustering of hierarchical clustering factors.

[0017] A third aspect of the present invention provides a computer-readable storage medium.

[0018] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the accident severity analysis method based on a hybrid linear model as described in the first aspect above.

[0019] A fourth aspect of the present invention provides a computer device.

[0020] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the steps in the accident severity analysis method based on a hybrid linear model as described in the first aspect above.

[0021] Compared with the prior art, the beneficial effects of the present invention are:

[0022] This invention comprehensively considers the relationship between various influencing factors and accident risks, reduces the impact of factors such as vehicles, roads, and the environment on the severity of traffic accidents, and clarifies the influence patterns of human factors. Only in this way can it provide theoretical support for safety education for drivers and passengers, promptly restrict unsafe behaviors of drivers and passengers, and improve the overall safety situation.

[0023] This invention finds that the number of vehicles involved in an accident, the type of accident, the vehicle type, and the lighting conditions have a significant impact on the severity of the accident, and exhibit hierarchical clustering, which can bias the research results.

[0024] This invention reduces the clustering of results by using a hybrid linear model and analyzes that among the characteristics of people in the vehicle, the driver's age, driver's gender, driver's physical condition, degree of injury, location of people, and safety measures have a significant impact on accidents. Attached Figure Description

[0025] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0026] Figure 1 This is a flowchart illustrating an accident severity analysis method based on a hybrid linear model, as shown in an embodiment of the present invention.

[0027] Figure 2This is a schematic diagram of multivariate correlation analysis shown in an embodiment of the present invention;

[0028] Figure 3 This is a comparison chart of the fitting statistics of the model before and after the present invention, as shown in an embodiment of the present invention. Detailed Implementation

[0029] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0030] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0031] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0032] It should be noted that the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of methods and systems according to various embodiments of this disclosure. It should be noted that each block in a flowchart or block diagram may represent a module, segment, or portion of code, which may include one or more executable instructions for implementing the logical functions specified in the various embodiments. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than that shown in the drawings. For example, two consecutively represented blocks may actually be executed substantially in parallel, or they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the flowcharts and / or block diagrams, and combinations of blocks in the flowcharts and / or block diagrams, may be implemented using a dedicated hardware-based system that performs the specified functions or operations, or using a combination of dedicated hardware and computer instructions.

[0033] Example 1

[0034] like Figure 1As shown, this embodiment provides a method for accident severity analysis based on a hybrid linear model. This embodiment uses the application of this method to a server as an example for illustration. It is understood that this method can also be applied to terminals, and can also be applied to systems including terminals, servers, and other components, and can be implemented through interaction between the terminal and the server. The server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network servers, cloud communication, middleware services, domain name services, CDN security services, and big data and artificial intelligence platforms. The terminal can be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, etc., but is not limited to these. The terminal and server can be directly or indirectly connected via wired or wireless communication, which is not limited herein. In this embodiment, the method includes the following steps:

[0035] Acquire personnel data, vehicle data, road data, and environmental data, and extract personnel characteristics, vehicle characteristics, road characteristics, and environmental characteristics;

[0036] Using a general linear model, we analyzed whether personnel characteristics, vehicle characteristics, road characteristics, and environmental characteristics have a significant impact on the severity of traffic accidents, and obtained hierarchical clustering factors with significance less than a set threshold.

[0037] By treating occupant characteristics as a fixed effect and hierarchical clustering factors as random effects, a mixed linear model is used to analyze the impact of occupant characteristics on the severity of traffic accidents, thereby reducing the clustering of hierarchical clustering factors.

[0038] The specific solution for this embodiment can be implemented by referring to the following content:

[0039] 1. Establishment of influencing factor dataset

[0040] 1.1 Accident Data Collection and Processing

[0041] This embodiment selects highway traffic accident data provided by the Highway Safety Information System (HSIS) website as the modeling object. The original dataset contains accident data from 2008 to 2012, including data tables related to people, vehicles, roads, and the environment. Because the feature fields of the personnel, vehicle, road, and environmental data are inconsistent, it is necessary to align accident numbers and standardize feature fields. The actual personnel, vehicle, road, and environmental data obtained are all statistically analyzed on an accident-by-accident basis.

[0042] This embodiment studies the impact of occupant characteristics on the severity of accidents. Specific criteria for classifying the degree of injury are shown in Table 1. Furthermore, the study integrated and statistically analyzed the casualties involved in each accident and classified the accident severity levels according to relevant regulations. Specific classification criteria are shown in Table 2.

[0043] 1.2 Extraction of Influencing Factors

[0044] After data screening and processing, a total of 8,650 accident data entries were obtained. The severity of the accident was used as the dependent variable, and 15 influencing factors were selected as independent variables from four aspects: people, vehicles, roads, and environment. These factors are: characteristics of the occupants (driver's age, driver's gender, driver's physical condition, driver's injury level, passenger's age, passenger's gender, passenger's injury level, occupant's position, and type of safety equipment), vehicle characteristics (accident type, vehicle type, and number of vehicles involved in the accident), road characteristics (road surface conditions), and environmental characteristics (weather and lighting conditions).

[0045] Table 1. Criteria for Classifying the Severity of Personnel Injuries

[0046]

[0047] Table 2. Criteria for Classifying Accident Severity

[0048]

[0049]

[0050] 1.3 Correlation Analysis of Influencing Factors

[0051] Considering that precise or high correlations may exist between variables in a linear regression model, potentially distorting the model estimate, further exploration of the correlations between variables is necessary. The correlation coefficient is a fundamental method used to test the correlation between variables; the Pearson correlation coefficient is a commonly used one. The Pearson correlation coefficient is an effective method for measuring the linear correlation between factors, and its calculation is simple and accurate. The Pearson correlation coefficient ranges from [-1, 1], and the larger the absolute value of the correlation coefficient, the stronger the linear correlation between the two variables. The results of this invention regarding multivariate correlation analysis are as follows: Figure 2 As shown in the figure. Related studies have shown that variable removal can be considered when the correlation coefficient is greater than 0.6. However, the multivariate correlation diagram of the accident causation index constructed in this invention shows that the correlation between the variables is relatively small. Therefore, variable removal is not required in this invention.

[0052] 2. Analysis Model of Factors Affecting Vehicle Occupants

[0053] 2.1 General Linear Model

[0054] Traditional simple linear models, such as Equation (1), require the assumptions of normality, independence, and homogeneity of variance to be satisfied when used. For stratified data with significant differences in dependent variable values, this can easily lead to excessive model errors, resulting in predictions that do not match the actual situation.

[0055] Y = αX + ε (1)

[0056] Where X represents the fixed effect, α is the influence coefficient of the fixed effect, and ε represents the random error.

[0057] 2.2 Mixed Linear Model

[0058] Compared to the traditional linear model, the mixed linear model retains only the first assumption, namely that the data follows a normal distribution, and introduces random effects parameters on the basis of the general linear model, as shown in equation (2).

[0059] Mixed linear models (MLMs) fully consider issues such as hierarchical structure of data, data clustering, and repeated measures data. They can accurately estimate parameters and test hypotheses for factors influencing the dependent variable, and further analyze the factors that lead to hierarchical clustering among data and factors that increase variability among individuals.

[0060] The stochastic intercept model (RIM) in mixed linear models is suitable for econometric data with hierarchical structure and clustered data properties.

[0061] The structure of the mixed linear model is as follows:

[0062] Y=αX+βZ+ε (2)

[0063] In the formula, Y represents the matrix vector of the measured values ​​of the dependent variable, X is the design matrix vector of the fixed-effect independent variable, α is the design matrix vector of the influence coefficient of the fixed-effect independent variable, Z is the design matrix vector of the random-effect independent variable, β is the design matrix vector of the influence coefficient of the random-effect independent variable, which follows a normal distribution with a mean vector of 0 and a variance / covariance matrix vector of M, and is represented as β~N(0,M). ε is the design matrix vector of the random error, which follows a normal distribution with a mean vector of 0 and a variance / covariance matrix vector of R, and is represented as ε~N(0,R).

[0064] Besides the characteristics of the occupants, vehicle characteristics, road characteristics, and environmental characteristics also influence the severity of traffic accidents. From an individual perspective, each accident is independent; however, from an analytical perspective, external conditions can cause similarities in accident characteristics. The interactions between accidents also lead to differences in severity under different conditions, thus exhibiting hierarchical clustering.

[0065] Because the data studied in this embodiment exhibits hierarchical clustering (vehicles, roads, and the environment form the second layer, while personnel form the first layer), although the accident severity level data is skewed, after logarithmic transformation and a normality test, a normally distributed dependent variable Y can be obtained. Therefore, it meets the application conditions of the Random Intercept Model (RIM). In practical applications, RIM involves two steps: establishing a null model to check whether the data truly possesses a hierarchical structure, and selecting fixed-effects and random-effects factors to finally establish an MLM analysis of influencing factors. The null model equation is as follows:

[0066] y ij =γ 00 +δ 0j +ε ij (3)

[0067] Among them, y ij This represents the outcome for individual i under condition j. γ 00 δ represents the total average or total intercept. 0j ε is a fixed parameter, representing a conditional level random variable that is unobservable or unobservable, and represents the distance from the intercept of the j-th condition to the total intercept. ij This is a random variable at the individual level, specifically the deviation of the i-th individual from the intercept of the group under the j-th condition. The null model does not contain any independent variable X; only the response variable Y and hierarchical grouping variables (vehicles, roads, and environment) are introduced during the modeling process. The statistical significance level is α = 0.05, meaning that when P < 0.05, it can be concluded that the data does indeed exhibit a hierarchical structure, and spatial clustering exists in the selected grouping variables for validation, making it suitable for MLM modeling analysis.

[0068] This embodiment uses the accident causation index set constructed as independent variables to model and explore the effects of individual and group-level factors on the dependent variable, forming a stochastic intercept model. The stochastic intercept model assumes that the intercept of the dependent variable varies across groups, but the regression slope is fixed for each group. Therefore, there is no interaction between factors at different levels. A simple two-level model expression is as follows:

[0069] y ij =(γ) 00 +γ 01 G 1j +γ 10 X 1ij )+(δ 0j +ε ij (4)

[0070] Such a two-level model is also called an intercept-only model. The model considers the dependent variable y... ij It is interpreted as a function of individual background and external environment. γ10 It is X 1ij The coefficient of γ represents the influence of individual background factors on the dependent variable, but its effect is independent of external circumstances; 01 It is an external coefficient, used as G 1j The direct functional expression of ε. ij The variance represents the magnitude of the variation of the dependent variable both internally and externally, which is not explained by the individuals and external factors included in the model. In actual modeling, a reference factor needs to be determined for the fixed-effects factor. A p-value < 0.05 is used as the statistical significance level; when p < 0.05, the fixed-effects factor or random-effects factor is considered to have an actual impact on the response variable Y, and the effect is determined to be positive or negative based on the parameter value. When evaluating the effect of a fixed-effects factor, the determined reference factor can be used as a reference.

[0071] 2.3 Establishment of Influencing Factor Analysis Model

[0072] This study investigates whether and how the characteristics of vehicle occupants affect the severity of road traffic accidents. However, even with the same occupant characteristics, different vehicles, roads, and environments can lead to different consequences. Therefore, a general linear model and a mixed linear model are used to verify whether vehicle, road, and environmental factors cause hierarchical clustering of results, and to analyze their impact on the severity of traffic accidents. To reduce the clustering effect among variables, hierarchical clustering factors are treated as random effects, while vehicle occupant characteristics are treated as fixed effects, and these are included in the mixed linear model to analyze the impact of vehicle occupant characteristics on the severity of traffic accidents, thereby improving the accuracy of the results.

[0073] 3. Results Analysis

[0074] 3.1 Clustering Factor Analysis

[0075] This embodiment first uses a general linear model to analyze whether attributes such as people, vehicles, roads, and the environment have a significant impact on the severity of traffic accidents. The model results are output using SPSS software, as shown in Table 3. Table 3 shows that accident type, vehicle type, number of vehicles involved in the accident, driver gender, driver injury severity, driver physical condition, passenger injury severity, passenger location, safety measures, and lighting conditions all have a significant impact on the severity of traffic accidents (P<0.05), while the effects of driver age, passenger age, passenger gender, weather, and road surface factors are not significant.

[0076] Table 3. Results of analysis using the general linear model

[0077]

[0078]

[0079] To further determine whether the analysis results exhibit clustering due to vehicle, road, and environmental variables, this embodiment uses a mixed linear model. The four variables in Table 3 with a significance level of P < 0.05—accident type, vehicle type, number of accident vehicles, and lighting conditions—are used as fixed effects inputs. The results are shown in Tables 4 and 5. Tables 4 and 5 show that the significance levels (P < 0.001) for the above four variables on accident severity, and their standard errors are not zero, thus indicating hierarchical clustering. Furthermore, the residual analysis results further demonstrate individual differences among traffic accidents. Therefore, using a general linear model that only considers fixed effects has certain limitations; a mixed linear model should be used to consider random effects to reduce clustering and conduct appropriate analysis.

[0080] 3.2 Analysis of occupant factors

[0081] Personnel characteristics were introduced as fixed effects into the mixed linear model, while four levels of clustering factors were introduced as random effects. The analysis results are as follows: Figure 3 As shown in Tables 6 and 7. Figure 3 It can be seen that the values ​​of -2 log-likelihood and AIC further decreased after introducing the random effect, indicating that introducing the random effect into this variable has a significant effect, and the mixed linear model has advantages in handling this type of data. The results of the fixed effects test of the model are shown in Table 6. Driver age, driver gender, driver physical condition, injury severity, personnel location, injury situation, and safety measures have a significant impact on the severity of the accident. The covariance test results in Table 7 show that the variance of the clustering factors further decreased compared with before the introduction of the random variable (see Table 5). The impact of accident type, vehicle type, number of accident vehicles, and lighting conditions is not significant, which also indicates that the introduction of the random effect significantly reduced the clustering of these four factors.

[0082] Table 4. Results of fixed-effects ANOVA for testing clustering.

[0083]

[0084]

[0085] Table 5. Results of random effects analysis for testing clustering.

[0086]

[0087] Table 6 Results of ANOVA with Fixed Effects

[0088]

[0089] Table 7 Results of random effects analysis

[0090]

[0091] The fixed-effects estimation results (Table 8) show that when the driver and passenger are teenagers aged 15-19, the significance levels are P = 0.038 and P = 0.047, respectively. This indicates that both have a significant impact on traffic accidents and are highly likely to increase the severity of accidents. For drivers of other age groups, this factor is not significant, and the negative effect on accidents becomes more pronounced with age, with drivers over 60 years old experiencing the lowest accident severity. When the driver is female, P < 0.001 and the estimated value is negative, indicating that women can significantly reduce the severity of accidents, suggesting that female drivers are more cautious than male drivers. Driving under the influence of medication, feeling unwell, or being drowsy also significantly increases the severity of accidents. Regarding the location of passengers, the back seat has a greater impact on the severity of traffic accidents. The positive impact of the driver's position is the greatest, easily exacerbating the severity of accidents. This indicates that passengers in this position are extremely prone to serious injury or death in an accident. The reason may be that compared to the front seats, the rear seats have fewer safety measures, and most rear passengers do not wear seat belts. Once a traffic accident occurs, the risk is extremely high. The driver's position has a significant negative effect on the severity of the accident, indicating that this position is the safest. The analysis results of the safety measures variable reflect that the use of safety measures such as seat belts, airbags, or child seats can effectively reduce the severity of injuries and fatalities. For motorcycle accidents, the severity of injuries is significantly reduced when the driver and passengers wear safety helmets, which is consistent with our daily traffic safety regulations. The more severe the injuries of the occupants, the greater and more significant the impact on the severity of the accident, which is also consistent with common sense.

[0092] 4. Conclusion

[0093] This embodiment, through the collation and merging of accident data, extracts features based on four aspects: people, vehicles, roads, and environment, analyzes and verifies the clustering of accidents, and establishes a hybrid linear model system to explore the impact of vehicle occupant characteristics on road traffic accidents. The conclusions are as follows:

[0094] (1) This embodiment found that the number of accident vehicles, accident type, vehicle type and lighting conditions have a significant impact on the severity of the accident, and have a hierarchical clustering effect, which may cause the research results to be biased.

[0095] (2) By using a mixed linear model to reduce the clustering of results, it was found that among the characteristics of people in the vehicle, the driver's age, driver's gender, driver's physical condition, degree of injury, location of people and safety measures have a significant impact on the accident.

[0096] (3) The risk of accidents is highest when both the driver and the passenger are teenagers aged 15 to 19. The risk of accidents is higher for male drivers than for female drivers. Drivers in poor physical condition will increase the risk of injury or death in accidents. The order of danger of seat position from highest to lowest is: rear seat > front passenger seat > driver's seat. Using safety measures such as seat belts, airbags, child seats or wearing safety helmets can effectively reduce the risk of fatal accidents.

[0097] The research findings will provide a theoretical basis for road safety training for drivers and passengers, and can serve as a reference for road management departments in formulating relevant policies and regulations. In preventing traffic accidents, it is essential to strengthen safety education for young drivers, emphasizing the importance of maintaining a positive mindset and physical condition while driving. Passengers are advised to sit in the front seat and strictly adhere to appropriate safety measures, such as wearing seatbelts and helmets, which will effectively improve driving safety.

[0098] Table 8. Results of Fixed Effects Analysis

[0099]

[0100]

[0101]

[0102] Example 2

[0103] This embodiment provides an accident severity analysis system based on a hybrid linear model.

[0104] An accident severity analysis system based on a hybrid linear model includes:

[0105] The data acquisition module is configured to acquire personnel data, vehicle data, road data, and environmental data, and extract personnel characteristics, vehicle characteristics, road characteristics, and environmental characteristics.

[0106] The general linear model processing module is configured to: use a general linear model to analyze whether personnel characteristics, vehicle characteristics, road characteristics, and environmental characteristics have a significant impact on the severity of traffic accidents, and obtain hierarchical clustering factors with significance less than a set threshold.

[0107] The mixed linear model processing module is configured to treat personnel characteristics as fixed effects and hierarchical clustering factors as random effects, and use a mixed linear model to analyze the impact of in-vehicle personnel characteristics on the severity of traffic accidents, thereby reducing the clustering of hierarchical clustering factors.

[0108] It should be noted that the data acquisition module, general linear model processing module, and mixed linear model processing module described above are the same examples and application scenarios implemented in the steps of Embodiment 1, but are not limited to the content disclosed in Embodiment 1. It should also be noted that the above modules, as part of the system, can be executed in a computer system such as a set of computer-executable instructions.

[0109] Example 3

[0110] This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps in the accident severity analysis method based on a hybrid linear model as described in Embodiment 1 above.

[0111] Example 4

[0112] This embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps in the accident severity analysis method based on a hybrid linear model as described in Embodiment 1 above.

[0113] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0114] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0115] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0116] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0117] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0118] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for analyzing accident severity based on a hybrid linear model, characterized in that, include: Acquire personnel data, vehicle data, road data, and environmental data, and extract personnel characteristics, vehicle characteristics, road characteristics, and environmental characteristics; Using a general linear model, we analyzed whether personnel characteristics, vehicle characteristics, road characteristics, and environmental characteristics have a significant impact on the severity of traffic accidents, and obtained hierarchical clustering factors with significance less than a set threshold. By treating personnel characteristics as a fixed effect and hierarchical clustering factors as random effects, a mixed linear model is used to analyze the impact of vehicle occupant characteristics on the severity of traffic accidents, thereby reducing the clustering of hierarchical clustering factors. The general linear model is as follows: ; The hybrid linear model is as follows: In the formula, A matrix vector representing the measured values ​​of the dependent variable. The design matrix vector for fixed effects independent variables. Design a matrix vector for the influence coefficients of the fixed-effects independent variables. The design matrix vector for the independent variables of random effects. Design a matrix vector for the influence coefficients of the random effect independent variables, following the principle that the mean vector is 0 and the variance / covariance matrix vector is... The normal distribution is expressed as: , Design a matrix vector for random error, which follows a mean vector of 0 and a variance / covariance matrix vector of... The normal distribution is expressed as ; The personnel characteristics, vehicle characteristics, road characteristics, and environmental characteristics meet the application conditions of the stochastic intercept model. The application of the stochastic intercept model includes: establishing a null model to detect whether the data truly has a hierarchical structure and selecting fixed-effects factors and random-effects factors, and finally establishing a mixed linear model to analyze the influencing factors. The null model is: ;in, Representative occurred at Under the conditions Individual outcomes Represents the overall average or overall intercept. It is a fixed parameter, a conditional random variable that is unobservable or unobservable, and represents the first... The distance from the intercept of each condition to the total intercept; It is a random variable at the individual level, that is, distributed at the th... The first condition The deviation of each individual from the intercept of the group; the null model contains no independent variables. The existence of the reaction variable means that only the reaction variable is introduced in the modeling process. and hierarchical grouping variables.

2. The accident severity analysis method based on a hybrid linear model according to claim 1, characterized in that, The hierarchical aggregation factors include: accident type, vehicle type, number of vehicles involved in the accident, and lighting conditions.

3. The accident severity analysis method based on a hybrid linear model according to claim 1, characterized in that, The personnel characteristics include driver's age, driver's gender, driver's physical condition, driver's injury level, passenger's age, passenger's gender, passenger's injury level, personnel location, and type of safety equipment; Vehicle characteristics include accident type, vehicle type, and number of vehicles involved in the accident; Road characteristics include road surface conditions; Environmental characteristics include weather and lighting conditions.

4. An accident severity analysis system based on a hybrid linear model, characterized in that, include: The data acquisition module is configured to acquire personnel data, vehicle data, road data, and environmental data, and extract personnel characteristics, vehicle characteristics, road characteristics, and environmental characteristics. The general linear model processing module is configured to: use a general linear model to analyze whether personnel characteristics, vehicle characteristics, road characteristics, and environmental characteristics have a significant impact on the severity of traffic accidents, and obtain hierarchical clustering factors with significance less than a set threshold. The mixed linear model processing module is configured to: treat personnel characteristics as fixed effects and hierarchical clustering factors as random effects, and use a mixed linear model to analyze the impact of vehicle occupant characteristics on the severity of traffic accidents, thereby reducing the clustering of hierarchical clustering factors; The general linear model is as follows: ; The hybrid linear model is as follows: In the formula, A matrix vector representing the measured values ​​of the dependent variable. The design matrix vector for fixed effects independent variables. Design a matrix vector for the influence coefficients of the fixed-effects independent variables. The design matrix vector for the independent variables of random effects. Design a matrix vector for the influence coefficients of the random effect independent variables, following the principle that the mean vector is 0 and the variance / covariance matrix vector is... The normal distribution is expressed as , Design a matrix vector for random error, which follows a mean vector of 0 and a variance / covariance matrix vector of... The normal distribution is expressed as ; The personnel characteristics, vehicle characteristics, road characteristics, and environmental characteristics meet the application conditions of the stochastic intercept model. The application of the stochastic intercept model includes: establishing a null model to detect whether the data truly has a hierarchical structure and selecting fixed-effects factors and random-effects factors, and finally establishing a mixed linear model to analyze the influencing factors. The null model is: ;in, Representative occurred at Under the conditions Individual outcomes Represents the overall average or overall intercept. It is a fixed parameter, a conditional random variable that is unobservable or unobservable, and represents the first... The distance from the intercept of each condition to the total intercept; It is a random variable at the individual level, that is, distributed at the th... The first condition The deviation of each individual from the intercept of the group; the null model contains no independent variables. The existence of the reaction variable means that only the reaction variable is introduced in the modeling process. and hierarchical grouping variables.

5. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in the accident severity analysis method based on a hybrid linear model as described in any one of claims 1-3.

6. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the accident severity analysis method based on a hybrid linear model as described in any one of claims 1-3.

Images