CN112349098A - Method for estimating accident severity by environmental elements in exit ramp area of expressway - Google Patents

Method for estimating accident severity by environmental elements in exit ramp area of expressway Download PDF

Info

Publication number
CN112349098A
CN112349098A CN202011208416.4A CN202011208416A CN112349098A CN 112349098 A CN112349098 A CN 112349098A CN 202011208416 A CN202011208416 A CN 202011208416A CN 112349098 A CN112349098 A CN 112349098A
Authority
CN
China
Prior art keywords
accident
severity
exit ramp
variable
independent variables
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202011208416.4A
Other languages
Chinese (zh)
Inventor
俞竞伟
黄开林
史荣珍
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing College of Information Technology
Original Assignee
Nanjing College of Information Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Nanjing College of Information Technology filed Critical Nanjing College of Information Technology
Priority to CN202011208416.4A priority Critical patent/CN112349098A/en
Publication of CN112349098A publication Critical patent/CN112349098A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0125—Traffic data processing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/243—Classification techniques relating to the number of classes
    • G06F18/24323—Tree-organised classifiers
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00—Computing arrangements using knowledge-based models
    • G06N5/01—Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/10—Services
    • G06Q50/26—Government or public services

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Tourism & Hospitality (AREA)
  • Development Economics (AREA)
  • Marketing (AREA)
  • Educational Administration (AREA)
  • General Business, Economics & Management (AREA)
  • Data Mining & Analysis (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Evolutionary Computation (AREA)
  • Quality & Reliability (AREA)
  • Operations Research (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Game Theory and Decision Science (AREA)
  • Chemical & Material Sciences (AREA)
  • Software Systems (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Analytical Chemistry (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Mathematical Physics (AREA)
  • Computing Systems (AREA)
  • Computational Linguistics (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Traffic Control Systems (AREA)

Abstract

The invention discloses a method for estimating the severity of an accident by environmental elements of an exit ramp area of an expressway, which comprises the steps of firstly, selecting accident data of the exit ramp of the expressway, selecting relevant environmental elements of the exit ramp area of the expressway as independent variables, integrating and screening, selecting complete accident data meeting requirements, establishing an accident sample database and training a decision tree; in the decision tree prediction model, sensitivity analysis is introduced to change the modelObserving the variation of each input variable and output result, regarding the output slope as the average variable influence of independent variable, predicting the non-binary discrete variable R of the linear fitting line of the result2Representing the nonlinear effect of the independent variables on the prediction results. The method can accurately and rigorously reflect the influence of the environmental elements in the exit ramp area of the expressway on the severity of the accident, is helpful for diagnosing the traffic safety problem of the exit ramp, and has important significance for improving the safety of the exit ramp and the whole expressway.

Description

Method for estimating accident severity by environmental elements in exit ramp area of expressway
Technical Field
The invention belongs to the field of traffic safety, and particularly relates to a method for estimating the severity of an accident by using environmental elements in an exit ramp area of a highway.
Background
The ramp area of the expressway is a zone where traffic accidents easily occur, and according to data statistics, the route length of the ramp accounts for less than 5% of the length of the whole expressway, but the traffic accidents occurring in the ramp area account for about 40% of the traffic accidents of the whole expressway, wherein the number of the traffic accidents occurring on the ramp area of the expressway is about 2 times that of the traffic accidents occurring on the ramp area of the expressway. The occurrence of traffic accidents is influenced by various factors such as people, vehicles, roads, environment and the like, and how to judge the influence of the environmental elements of the exit ramp of the expressway on the severity of the accidents is paid attention by a plurality of traffic safety management and researchers. The factors which are greatly related to the severity of the traffic accident are judged, the diagnosis of the traffic safety problem existing in the exit ramp can be facilitated, a basis is provided for the basic design and planning of road traffic, and the method has important significance for improving the safety of the exit ramp and the whole expressway.
At present, a decision tree model is applied to accident severity prediction, and the decision tree prediction model does not need to determine the linear relation between independent variables and dependent variables, so that the model has higher prediction accuracy than a traditional statistical model. But their main drawback is that they operate internally, usually like a "black box", and do not directly output the correlation between accident severity and independent variables, and this method cannot be used to explore the non-linear relationship between variables. Some researchers have attempted to introduce sensitivity analysis into decision tree prediction to explore the effect of independent variables on the prediction results, but they have only calculated the average change of the output results under the condition of input independent variable variation and cannot explore the nonlinear effect of the independent variables on the prediction results. And the existing research on the influence factors of the traffic accidents mainly aims at the overall condition of the expressway and lacks accident analysis aiming at an exit ramp area. The traffic environment of the exit ramp area of the expressway is complex, the traffic design and operation are unique, the traffic safety influence factors are many, the accident occurrence reason and the severity degree are greatly different from those of other road sections, and therefore, the environmental factors influencing the accident severity degree of the exit ramp need to be specially researched.
Disclosure of Invention
The purpose of the invention is as follows: aiming at the defects of the prior art, the invention provides a method for estimating the accident severity by environmental elements in an exit ramp area of an expressway, and the nonlinear influence of the environmental elements in the exit ramp area of the expressway on the accident severity is researched.
The technical scheme is as follows: the invention discloses a method for estimating the severity of an accident by environmental elements in an exit ramp area of a highway, which comprises the following steps of:
(1) acquiring accident data of an exit ramp area of a highway in advance;
(2) establishing an accident sample data set of an exit ramp area of the expressway, and selecting environmental elements as independent variables;
(3) dividing the data sorted in the step (2) into a plurality of grades according to the accident severity, and sorting the data according to the injury severity from no injury to death according to a division criterion;
(4) constructing a data structure of a decision tree, establishing a prediction model of the accident severity of the expressway exit ramp area based on the decision tree, and training the model;
(5) analyzing the correlation between the independent variable and the accident severity by using sensitivity, predicting the nonlinear influence of the environmental elements of the exit ramp of the expressway on the accident severity, and revealing the influence of the independent variable on the accident severity;
(6) evaluating the nonlinear influence of the environmental factors of the expressway exit ramp area based on the decision tree prediction model on the severity of the accident: and (5) calculating sensitivity analysis evaluation indexes of all independent variables according to the step (5), and comparing the influence of environmental elements of the exit ramp area of the high-speed road on the severity of the accident.
Further, the environment elements in the step (2) include exit ramp type, number of main lines, number of exit ramp lanes, length of deceleration lane, total length of ramp, road surface type, road surface condition, shoulder shape, shoulder width, rear speed limit of main line, speed limit difference between main line and exit ramp, lighting condition, weather condition, land use type, average daily traffic volume of main line, average daily traffic volume of exit ramp, presence or absence of drunk driving, poison driving and accident type.
Further, the step (2) comprises the steps of:
(21) selecting environmental elements from traffic accident data samples in the expressway exit ramp area as independent variables, wherein the content covered by the elements comprises geometric characteristics, meteorological conditions, lighting conditions and traffic flow characteristics of accident areas;
(22) normalizing the accident data acquired in the step (1), performing linear conversion on the accident data, adopting minimum-maximum normalization, mapping the result to [0,1], and calculating the following formula:
Figure BDA0002757902470000021
(23) and integrating and screening the data according to the characteristics of the required independent variables, and selecting complete accident samples meeting requirements to establish an accident sample database.
Further, the step (4) comprises the steps of:
(41) classifying in a decision tree by using a structure similar to a flow chart, wherein each internal node represents the test of an independent variable, each branch represents the test result, each leaf node represents a class label, and a path from a root to a leaf represents a classification rule;
(42) for the accident severity prediction model, each node in the decision tree represents an accident severity predictor, each branch represents a state of a feature of the predictor, and the leaves represent the expected accident severity, which depends on the independent variables and the accident severity information provided in the accident sample set;
(43) when a new sample of the incident in the test data set is obtained, a prediction is made as to the severity of the incident by following the path from the root to the leaves in the tree using the partition eigenvalues.
Further, the step (5) includes the steps of:
(51) changing the selected independent variables one by adopting a successive item substitution method, namely manually executing small disturbance on one of the independent variables, fixing other independent variables to be unchanged, and calculating the variation of sensitivity analysis indexes under all different values for discrete independent variables such as binary and multivariate variables; keeping other independent variables fixed for continuous independent variables, setting the independent variables as minimum values, increasing the independent variables by 10% each time until the independent variables reach the maximum values, and calculating the change of the sensitivity analysis index;
(52) the sensitivity analysis index adopts the slope of a fit line of an output result and a non-binary discrete variable R of the fit line2In terms of scale, the slope is taken as the average variable influence of the independent variable to predict the non-binary discrete variable R of the straight line fitting line of the result2Exploring the nonlinear effects of the independent variables on the prediction results:
the slope of the line is determined by a least squares method using the accident sample data to provide a slope value which is calculated as follows:
Figure BDA0002757902470000031
wherein x isiThe value of the independent variable for the ith observed accident; y isiThe value of the dependent variable for the ith observed accident sample;
Figure BDA0002757902470000032
is the average of the independent variables;
Figure BDA0002757902470000033
is the average value of the dependent variable; n is the total accident data sample size;
if the influence of the independent variable on the severity of the accident is positive correlation, the larger the independent variable value is, the more the traffic accident is, and vice versa; the greater the variable impact value, i.e., the greater the slope, the greater the impact on crash severity;
non of straight line fitBinary discrete variable R2Calculated according to the following formula:
Figure BDA0002757902470000034
wherein the content of the first and second substances,
Figure BDA0002757902470000035
is a weighted average of the dependent variables of the ith observation sample; r2Whether the output changes linearly under the condition of different variable variables of the input is measured, R2The larger the linear variation of the variable, the closer the non-linearity of the effect of the independent variable on the severity of the accident.
Has the advantages that: compared with the prior art, the invention has the beneficial effects that: 1. the method comprises the steps of constructing a decision tree prediction model, analyzing traffic accidents in the expressway exit ramp area with definite results, searching for characteristics in data, selecting environmental elements affecting the expressway exit ramp as independent variables according to actual conditions, dividing accident severity, establishing an accident sample database and training a decision tree; 2. the method is used for carrying out sensitivity analysis on the input variables of the decision tree learning model and selecting the slope of a linear fit line of an output result and the non-binary discrete variable R2Evaluating the influence of environmental factors of an exit ramp area of the expressway on the severity of the accident; compared with the independent variable sensitivity analysis applied to a machine learning model, the method disclosed by the invention is more accurate and strict in accordance with the nonlinear influence estimation of the environmental elements in the actual expressway exit ramp area on the accident severity, and the influence of each environmental element on the accident severity can be known by applying the technology, so that a basis is provided for improving the geometric design and traffic management control of the expressway exit ramp, and the method has an important significance for improving the driving safety of the exit ramp and even the whole expressway.
Drawings
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a schematic diagram of input and output variables of a decision tree model;
FIG. 3 is a flow chart of decision tree prediction model generation;
FIG. 4 is a diagram illustrating the calculation result of the slope of the independent variable influence.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings.
The invention provides a method for estimating the severity of an accident by environmental elements in an exit ramp area of a highway, which specifically comprises the following steps as shown in figure 1:
step 1: and acquiring accident data of the exit ramp area of the expressway in advance.
In this embodiment, a total of 5538 incidents were analyzed, with data from three years of incident statistics in florida, usa. The accident statistics area includes 326 highway sections in total, and the research area range is a deceleration area and a ramp exit area with the length of 762 meters (2500 feet).
Step 2: and processing data, establishing an accident sample data set of an exit ramp area of the expressway, and selecting environmental elements as independent variables.
Selecting environmental elements from traffic accident data samples of the expressway exit ramp area as independent variables, wherein the content covered by the elements comprises the following steps: geometric characteristics of the accident area, weather conditions, lighting conditions, traffic flow characteristics, and the like. The covered variable attributes can be different, such as land use type, road surface type, shoulder type, lighting condition, road surface condition, weather condition, whether the drunk driving/poisonous driving is related or not is a binary discrete variable; the exit ramp type, the number of main line lanes and the number of exit ramp lanes are multivariate variables; the deceleration lane length, total ramp length, right shoulder width, main line rear speed limit, and speed limit difference between the main line and the exit ramp are continuous type variables.
Performing linear conversion on the original data of the accidents in the exit ramp area of the expressway collected in the step 1, mapping the result to [0,1] by adopting Min-Max Normalization, and calculating the following formula:
Figure BDA0002757902470000051
integrating and screening data according to the characteristics of the required independent variables, and selecting complete accident samples meeting requirements to establish an accident sample database; the normalized data was classified by the 18 selected independent variables as shown in table 1.
TABLE 1 independent variable Classification statistics
Figure BDA0002757902470000052
Figure BDA0002757902470000061
The land use type, the road surface type, the shoulder type, the lighting condition, the road surface dry and wet condition, the weather condition and the existence of drunk driving/poisonous driving are binary discrete variables; the exit ramp type, the number of main line lanes and the number of exit ramp lanes are multivariate variables; the deceleration lane length, total ramp length, right shoulder width, main line rear speed limit and speed limit difference between the main line and the exit ramp are continuous variables;
and step 3: and (3) dividing accident severity grades: and D, dividing the data sorted in the step two into grades according to the severity of the accident, wherein the division criterion is the serious injury condition, and classifying and counting the accident samples by dividing five types of severity as shown in the table 2.
TABLE 2 Accident severity level and frequency statistics
Figure BDA0002757902470000062
And 4, step 4: and constructing a data structure of a decision tree, establishing a prediction model of the severity of the accident in the exit ramp area of the expressway based on the decision tree, and training the model, wherein a schematic diagram of input and output variables of the decision tree model is shown in FIG. 2.
And (3) selecting characteristics according to the principle that the purities of all data subsets after the accident sample set is divided are higher than those of the accident sample set before the division by the highway exit ramp traffic accident influence factors, wherein the purities of the subsets are measured by an information gain rate, 18 independent variables of the selected highway exit ramps are used as candidate characteristics, the characteristics that the information gain is higher than the average level are found out from the candidate characteristics, and then the characteristics with the highest information gain rate are selected from the characteristics.
A structure similar to the flow chart is used for classification in the decision tree. Each internal node represents a test on an argument, each branch represents the result of the test, each leaf node represents a class label, and the path from the root to the leaf represents a classification rule. For the incident severity prediction model, each node in the decision tree represents an incident severity predictor, and each branch represents a state of a feature of the predictor. The leaves (or terminal nodes) represent the expected severity of the incident, which depends on the independent variables and incident severity information provided in the incident sample set. When a new sample of the incident in the test data set is obtained, a prediction can be made as to the severity of the incident by following the path from the root to the leaves in the tree using the partition eigenvalues. More preferably, the key to building a decision tree is how to select the best attributes. The fault samples in the branch nodes belong to the same category as much as possible, i.e. the purity of the node should be highest. And leaf nodes which do not belong to the sum of the weights in the category to which the leaf nodes belong and branch nodes which are the sum of all the error leaf nodes are cut off in the process of generating the decision tree.
And 5: carrying out independent variable sensitivity analysis of the accident prediction model: and analyzing the correlation between the independent variable and the accident severity by using sensitivity, predicting the nonlinear influence of the environmental elements of the exit ramp of the expressway on the accident severity, and revealing the influence of the independent variable on the accident severity.
Changing the selected independent variables one by adopting a successive substitution method, namely manually executing small disturbance on one independent variable, fixing other independent variables to be unchanged, and calculating the change of sensitivity analysis indexes under all different values for discrete independent variables such as binary and multivariate variable land use types, road surface types, shoulder types, lighting conditions, road surface dry and wet conditions, weather conditions, whether or not the discrete independent variables relate to drunk driving/toxic driving, rear-end collision accidents, scraping accidents and side collision, exit ramp types, main line lane numbers and exit ramp lane numbers; keeping other independent variables fixed for continuous independent variables such as the length of a deceleration lane, the total length of ramps, the width of right shoulders, the rear speed limit of a main line and the speed limit difference between the main line and an exit ramp, setting the independent variables as minimum values, increasing the independent variables by 10% each time until the independent variables reach the maximum values, and calculating the change of a sensitivity analysis index;
the sensitivity analysis index adopts the slope of a fit line of an output result and a non-binary discrete variable R of the fit line2In terms of scale, the slope is taken as the average variable influence of the independent variable to predict the non-binary discrete variable R of the straight line fitting line of the result2And (3) exploring the nonlinear influence of the independent variable on the prediction result, and specifically calculating as follows:
the slope of the line is determined by the least squares method, and the calculation formula is as follows:
Figure BDA0002757902470000071
wherein x isiThe value of an independent variable of the ith observation; y isiThe value of the dependent variable of the ith observation data;
Figure BDA0002757902470000072
is the average of the independent variables;
Figure BDA0002757902470000073
is the average value of the dependent variable; n is the total data volume;
taking the variable deceleration lane length as an example, the slope calculation result can be represented by fig. 3. Each line has a formula of the same form: y-ax + b, where a represents slope (mean variable influence). The positive and negative values of the slope value reflect the influence trend of the independent variable. If the independent variable effect is positive, a larger value of the independent variable will result in a greater severity of the accident and vice versa. And the greater the variable impact, the greater its impact on the severity of the traffic accident. Fig. 3 shows that the average variable impact of the deceleration lane length on the property loss only incidents in the output results in the decision tree model is greatest.
To evaluate the non-linearity of variable impact on accident severity, a non-binary discrete variable R of a line of linear fit is calculated2。
Figure BDA0002757902470000081
Wherein the content of the first and second substances,
Figure BDA0002757902470000082
is a weighted average of the dependent variables of the ith observation.
R2The larger the value, the closer the linear change of the variable. In this example, multivariate variables such as exit ramp type, number of main line lanes, and number of exit ramp lanes are calculated, respectively; nonlinear influence index R of continuous variable deceleration lane length, total ramp length, right shoulder width, main line rear speed limit and speed limit difference between main line and exit ramp on accident severity2And comparing the calculation results. Taking the variable deceleration lane length as an example, R2The results of the calculations are shown in FIG. 4, which shows that the non-linear effect of the deceleration lane length on the loss of property only accident is greatest.
Step 6: evaluating the nonlinear influence of the environmental factors of the expressway exit ramp area based on the decision tree prediction model on the severity of the accident: and (5) calculating sensitivity analysis evaluation indexes of all independent variables according to the step 5, and comparing the influence of environmental elements of the exit ramp area of the high-speed highway on the severity of the accident.
In the embodiment, the nonlinear influence estimation of the 18 selected environmental influence factors on the accident severity is used to conclude that the linear influence of the speed difference on the accident severity is the minimum, and the linear influence of the number of ramp lanes on the accident severity is the maximum.
According to the method, the nonlinear influence of the environmental factors of the exit ramp area of the expressway on the accident severity is researched, in the technology, the environmental factors influencing the accident severity of the exit ramp area of the expressway are selected as independent variables according to experience, accident data are classified, the accident severity is divided, and the division criterion is the serious injury condition. An accident severity prediction model is established based on a decision tree algorithm, the sensitivity is used for analyzing the correlation between the independent variable and the accident severity, and the nonlinear influence of the environmental elements of the exit ramp of the expressway on the accident severity is predicted so as to reveal the influence of the environmental influence factor variable on the accident severity.

Claims (5)

1. A method for estimating the severity of an accident by environmental elements in an exit ramp area of a highway is characterized by comprising the following steps of:
(1) acquiring accident data of an exit ramp area of a highway in advance;
(2) establishing an accident sample data set of an exit ramp area of the expressway, and selecting environmental elements as independent variables;
(3) dividing the data sorted in the step (2) into a plurality of grades according to the accident severity, and sorting the data according to the injury severity from no injury to death according to a division criterion;
(4) constructing a data structure of a decision tree, establishing a prediction model of the accident severity of the expressway exit ramp area based on the decision tree, and training the model;
(5) analyzing the correlation between the independent variable and the accident severity by using sensitivity, predicting the nonlinear influence of the environmental elements of the exit ramp of the expressway on the accident severity, and revealing the influence of the independent variable on the accident severity;
(6) evaluating the nonlinear influence of the environmental factors of the expressway exit ramp area based on the decision tree prediction model on the severity of the accident: and (5) calculating sensitivity analysis evaluation indexes of all independent variables according to the step (5), and comparing the influence of environmental elements of the exit ramp area of the high-speed road on the severity of the accident.
2. The method according to claim 1, wherein the environmental elements in step (2) include exit ramp type, number of main lanes, number of exit ramp lanes, deceleration lane length, total ramp length, road surface type, road surface condition, shoulder shape, shoulder width, main line rear speed limit, speed limit difference between main line and exit ramp, lighting condition, weather condition, land use type, main line average daily traffic volume, exit ramp average daily traffic volume, presence or absence of drunk driving, poison driving, and accident type.
3. The method for estimating the severity of an accident based on environmental elements of an off-ramp highway area according to claim 1, wherein said step (2) comprises the steps of:
(21) selecting environmental elements from traffic accident data samples in the expressway exit ramp area as independent variables, wherein the content covered by the elements comprises geometric characteristics, meteorological conditions, lighting conditions and traffic flow characteristics of accident areas;
(22) normalizing the accident data acquired in the step (1), performing linear conversion on the accident data, adopting minimum-maximum normalization, mapping the result to [0,1], and calculating the following formula:
Figure FDA0002757902460000011
(23) and integrating and screening the data according to the characteristics of the required independent variables, and selecting complete accident samples meeting requirements to establish an accident sample database.
4. The method for estimating the severity of an accident based on environmental elements of an exit ramp area of a highway according to claim 1, wherein the step (4) comprises the steps of:
(41) classifying in a decision tree by using a structure similar to a flow chart, wherein each internal node represents the test of an independent variable, each branch represents the test result, each leaf node represents a class label, and a path from a root to a leaf represents a classification rule;
(42) for the accident severity prediction model, each node in the decision tree represents an accident severity predictor, each branch represents a state of a feature of the predictor, and the leaves represent the expected accident severity, which depends on the independent variables and the accident severity information provided in the accident sample set;
(43) when a new sample of the incident in the test data set is obtained, a prediction is made as to the severity of the incident by following the path from the root to the leaves in the tree using the partition eigenvalues.
5. The method for estimating the severity of an accident based on environmental elements of an off-ramp highway area according to claim 1, wherein said step (5) comprises the steps of:
(51) changing the selected independent variables one by adopting a successive item substitution method, namely manually executing small disturbance on one of the independent variables, fixing other independent variables to be unchanged, and calculating the variation of sensitivity analysis indexes under all different values for discrete independent variables such as binary and multivariate variables; keeping other independent variables fixed for continuous independent variables, setting the independent variables as minimum values, increasing the independent variables by 10% each time until the independent variables reach the maximum values, and calculating the change of the sensitivity analysis index;
(52) the sensitivity analysis index adopts the slope of a fit line of an output result and a non-binary discrete variable R of the fit line2In terms of scale, the slope is taken as the average variable influence of the independent variable to predict the non-binary discrete variable R of the straight line fitting line of the result2Exploring the nonlinear effects of the independent variables on the prediction results:
the slope of the line is determined by a least squares method using the accident sample data to provide a slope value which is calculated as follows:
Figure FDA0002757902460000021
wherein x isiThe value of the independent variable for the ith observed accident; y isiThe value of the dependent variable for the ith observed accident sample;
Figure FDA0002757902460000022
is the average of the independent variables;
Figure FDA0002757902460000023
is the average value of the dependent variable; n is the total accident data sample size;
if the influence of the independent variable on the severity of the accident is positive correlation, the larger the independent variable value is, the more the traffic accident is, and vice versa; the greater the variable impact value, i.e., the greater the slope, the greater the impact on crash severity;
non-binary discrete variable R of straight line fitting line2Calculated according to the following formula:
Figure FDA0002757902460000031
wherein the content of the first and second substances,
Figure FDA0002757902460000032
is a weighted average of the dependent variables of the ith observation sample; r2Whether the output changes linearly under the condition of different variable variables of the input is measured, R2The larger the linear variation of the variable, the closer the non-linearity of the effect of the independent variable on the severity of the accident.
CN202011208416.4A 2020-11-03 2020-11-03 Method for estimating accident severity by environmental elements in exit ramp area of expressway Pending CN112349098A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011208416.4A CN112349098A (en) 2020-11-03 2020-11-03 Method for estimating accident severity by environmental elements in exit ramp area of expressway

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011208416.4A CN112349098A (en) 2020-11-03 2020-11-03 Method for estimating accident severity by environmental elements in exit ramp area of expressway

Publications (1)

Publication Number Publication Date
CN112349098A true CN112349098A (en) 2021-02-09

Family

ID=74356130

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011208416.4A Pending CN112349098A (en) 2020-11-03 2020-11-03 Method for estimating accident severity by environmental elements in exit ramp area of expressway

Country Status (1)

Country Link
CN (1) CN112349098A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113611114A (en) * 2021-08-03 2021-11-05 长安大学 Urban expressway traffic accident meteorological factor sensitivity analysis method

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110458244A (en) * 2019-08-20 2019-11-15 合肥工业大学 A kind of traffic accident Severity forecasting method applied to Regional Road Network
CN110555565A (en) * 2019-09-09 2019-12-10 南京东控智能交通研究院有限公司 Decision tree model-based expressway exit ramp accident severity prediction method

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110458244A (en) * 2019-08-20 2019-11-15 合肥工业大学 A kind of traffic accident Severity forecasting method applied to Regional Road Network
CN110555565A (en) * 2019-09-09 2019-12-10 南京东控智能交通研究院有限公司 Decision tree model-based expressway exit ramp accident severity prediction method

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113611114A (en) * 2021-08-03 2021-11-05 长安大学 Urban expressway traffic accident meteorological factor sensitivity analysis method

Similar Documents

Publication Publication Date Title
CN109993223B (en) Pavement use performance prediction method, storage medium and electronic equipment
CN114783183B (en) Traffic situation algorithm-based monitoring method and system
CN110503245B (en) Prediction method for large-area delay risk of airport flight
CN112506990B (en) Hydrological data anomaly detection method based on spatiotemporal information
CN110796859A (en) Real-time traffic state identification and accident risk early warning method based on traffic flow
CN110555565A (en) Decision tree model-based expressway exit ramp accident severity prediction method
CN104835103A (en) Mobile network health evaluation method based on neural network and fuzzy comprehensive evaluation
Pérez-Guillén et al. Data-driven automated predictions of the avalanche danger level for dry-snow conditions in Switzerland
CN110910641A (en) Neural network-based real-time risk assessment method for highway traffic operation
CN110674858B (en) Traffic public opinion detection method based on space-time correlation and big data mining
CN111126868B (en) Road traffic accident occurrence risk determination method and system
CN113496314B (en) Method for predicting road traffic flow by neural network model
CN112149922A (en) Method for predicting severity of accident in exit and entrance area of down-link of highway tunnel
CN114299742B (en) Speed limit information dynamic identification and update recommendation method for expressway
CN116524712A (en) Highway congestion prediction method, system and device integrating space-time associated data
CN105678406A (en) Short-term load prediction method based on cloud model
CN116245367A (en) Dangerous truck transportation risk assessment method and system based on hierarchical fuzzy neural network
CN115948964A (en) Road flatness prediction method based on GA-BP neural network
Li et al. New model of travel-time prediction considering weather conditions: case study of urban expressway
CN112349098A (en) Method for estimating accident severity by environmental elements in exit ramp area of expressway
AlKheder et al. Urban traffic prediction using metrological data with fuzzy logic, long short-term memory (LSTM), and decision trees (DTs)
Pecherková et al. Analysis of discrete data from traffic accidents
CN112541636A (en) Power transmission line icing risk early warning method and device, medium and electronic equipment
CN116756825A (en) Group structural performance prediction system for middle-small span bridge
CN111737993A (en) Method for extracting health state of equipment from fault defect text of power distribution network equipment

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication

Application publication date: 20210209

RJ01 Rejection of invention patent application after publication