WO2021027153A1

WO2021027153A1 - Method and apparatus for constructing traffic flow data analysis model

Info

Publication number: WO2021027153A1
Application number: PCT/CN2019/117989
Authority: WO
Inventors: 杜艳艳
Original assignee: 平安科技（深圳）有限公司
Priority date: 2019-08-15
Filing date: 2019-11-13
Publication date: 2021-02-18
Also published as: CN110444022A

Abstract

A method and apparatus for constructing a traffic flow data analysis model, relating to the technical field of data processing. The method comprises: establishing a data channel with a traffic flow data monitoring system, determining a corresponding road segment according to a target area, and extracting traffic flow data of the target area from the data channel (S110); obtaining historical traffic flow data in different sub-periods of a set period of each road segment as a training sample (S120); updating the training sample of each road segment according to a set maximum number of iterations, and obtaining an optimal fitness value of all road segments using a shuffled frog leaping algorithm (S130); and obtaining a connection weight value using the optimal fitness value of all road segments, and constructing, according to the connection weight value, a traffic flow data analysis model of a wavelet neural network (S140). The method fills a long-standing gap in analysis tools for traffic flow data, and improves the efficiency of traffic flow data prediction.

Description

Method and device for constructing traffic flow data analysis model

This application claims the priority of a Chinese patent application filed with the Chinese Patent Office on August 15, 2019, the application number is 201910754650.8, and the application title is "Method and Device for Constructing Traffic Flow Data Analysis Model", the entire content of which is incorporated by reference In this application.

Technical field

This application relates to the technical field of data processing. Specifically, this application relates to a method and device for constructing a traffic flow data analysis model.

Background technique

With the development of urban transportation networks, traffic flow is more susceptible to many factors, and the characteristics of randomness are becoming more and more prominent. At present, in response to this problem, an intelligent transportation system is used to record historical traffic flow data in the same time period in the near future, and to predict the recent traffic flow based on the recorded content. Although this method can predict real-time traffic flow data to a certain extent, there is no tool that can be widely applied to long-term road condition analysis and does not help long-term traffic flow forecasting.

Summary of the invention

In order to overcome the above technical problems, especially the prior art can only obtain historical traffic flow data from the same time period in a short period of time, and cannot well solve the problem of predicting long-term traffic flow data, the following technical solutions are proposed:

In the first aspect, this application provides a method for constructing a traffic flow data analysis model, which includes the following steps:

Establish a data channel with the traffic flow data monitoring system, determine the corresponding road section according to the target area, and extract the traffic flow data of the target area from the data channel;

Obtain historical traffic flow data of different sub-time periods of each road section in the set time period as training samples;

According to the set maximum number of iterations, the training samples of each road section are updated, and the hybrid frog leaping algorithm is used to obtain the optimal fitness value of all road sections;

The optimal fitness values of all road sections are used to obtain connection weights, and a wavelet neural network traffic flow data analysis model is constructed according to the connection weights.

In the second aspect, this application also provides a device for constructing a traffic flow data analysis model, which includes:

The extraction module is used to establish a data channel with the traffic flow data monitoring system, determine the corresponding road section according to the target area, and extract the traffic flow data of the target area from the data channel;

The training sample setting module is used to obtain the historical traffic flow data of each road section in different sub-time periods of the set time period as training samples;

The iterative calculation module is used to update the training samples of each road section according to the set maximum number of iterations, and obtain the optimal fitness value of all road sections by using the hybrid leapfrog algorithm;

The construction module is used for obtaining connection weights by using the optimal fitness values of all road sections, and constructing and obtaining a traffic flow data analysis model of the wavelet neural network according to the connection weights.

In the third aspect, this application also provides a server, which includes:

One or more processors;

Memory

One or more computer programs, wherein the one or more computer programs are stored in the memory and configured to be executed by the one or more processors, and the one or more computer programs are configured to execute The method for constructing a traffic flow data analysis model described in the embodiment of the above first aspect.

In a fourth aspect, the present application also provides a computer non-volatile readable storage medium, the computer non-volatile readable storage medium stores a computer program, and when the computer program is executed by a processor, the first aspect is realized The method for constructing the traffic flow data analysis model described in the embodiment.

The method and device for constructing a traffic flow data analysis model provided by this application obtain traffic flow data of all road sections in a target area through a data channel with a traffic flow data monitoring system; and according to different sub-time periods of the set time period The historical traffic flow data is used as the training sample, and the optimal fitness value is obtained using the hybrid leapfrog algorithm, and the traffic flow data analysis model of the wavelet neural network is obtained.

On this basis, this application also further technically optimizes the technical solution of the construction method and device of the traffic flow data analysis model, and divides the traffic flow directions of all road sections in the target area into the injection direction and the drainage direction, The hybrid frog leaping algorithm is used to train the injected traffic flow and the diverted traffic flow of each road section to obtain the corresponding traffic flow data analysis model. The traffic flow data analysis model separately predicts the injected traffic flow and the diversion traffic flow, and can obtain the prediction results of the traffic flow of all road sections in the target area more accurately.

The technical solution provided in this application uses the hybrid frog leaping algorithm to obtain the connection weights of the wavelet neural network prediction model. It uses the characteristics of the hybrid frog leaping algorithm with strong global search ability, fast convergence speed, and less parameter configuration, making it possible to Fast convergence to obtain the global optimal solution, so the use of the initial wavelet neural network parameters obtained by the hybrid frog leaping algorithm to train the traffic flow data analysis model can achieve the goal of improving the prediction speed and prediction accuracy when predicting the traffic flow. At the same time, the method and device for constructing a traffic flow data analysis model provided by this application have strong global search capabilities and can be well adapted to different traffic flow data samples after updating and iteration, and therefore can be adapted to different traffic flow data samples in different periods and periods. Flow data constructs a corresponding traffic flow data analysis model, filling the gap of long-term traffic flow data analysis tools.

The additional aspects and advantages of this application will be partly given in the following description, which will become obvious from the following description, or be understood through the practice of this application.

Description of the drawings

The above and/or additional aspects and advantages of the present application will become obvious and easy to understand from the following description of the embodiments in conjunction with the accompanying drawings, in which:

Fig. 1 is a flowchart of a method for constructing a traffic flow data analysis model according to an embodiment of the application;

2 is a flowchart of a method for constructing a traffic flow data analysis model according to another embodiment of the application;

FIG. 3 is a schematic diagram of a device for constructing a traffic flow data analysis model according to an embodiment of the application;

FIG. 4 is a schematic structural diagram of a server according to an embodiment of the application.

detailed description

The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings.

The singular forms "a", "an", "said" and "the" used herein may also include plural forms. The term "comprising" used in the specification of this application refers to the presence of the described features, integers, steps, operations, elements, and/or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, Elements, components and/or groups of them. It should be understood that the term "and/or" used herein includes all or any unit and all combinations of one or more associated listed items.

Although this method can predict real-time traffic flow data to a certain extent, there is no one that can be widely applied to long-term road condition analysis models and does not help long-term traffic flow forecasting.

Currently, there is a lack of long-term and widely applicable road condition analysis tools. This application provides a method for constructing a traffic flow data analysis model. Please refer to FIG. 1, which is an embodiment of the construction of a traffic flow data analysis model. The flow chart of the method includes the following steps:

S110: Establish a data channel with the traffic flow data monitoring system, determine a corresponding road section according to the target area, and extract traffic flow data of the target area from the data channel.

In this application, the target area is an area including several road sections. The road section is a general traffic road section, which must form a direct or indirect connection with other road sections, and will interact with the traffic flow data of all road sections, and act on the traffic flow data of other road sections connected to it, especially A road section that directly inputs or outputs a traffic flow to the target road section.

The acquisition of traffic flow data for all road sections is completed by the traffic flow data monitoring system. When acquiring traffic flow data of a corresponding road section, first determine the road section corresponding to the target area, and add a label to the road section. According to the label of the road section, the traffic flow data of the corresponding road section is obtained through the data channel established with the traffic flow data monitoring system.

S120. Obtain historical traffic flow data of different sub-time periods in a set time period for each road section as a training sample.

In this embodiment, the evaluation time period of the traffic flow data of all road sections in the target area is determined as a set time period, and the evaluation time period is divided into several sub-time periods with equal time intervals. For each road section in the target area, the traffic flow data in each sub-time period is obtained, and part of the traffic flow data is used as the training sample x _it .

S130: According to the set maximum number of iterations, the training samples of each road section are updated, and the optimal fitness value of all road sections is obtained by using the hybrid leaping algorithm.

For this step, the hybrid leapfrog algorithm is used to calculate the fitness value of the traffic flow data of each road section to obtain the optimal fitness value, and the traffic flow data analysis model is trained.

In the process of using the hybrid leaping algorithm for training, the training samples obtained in step S120 are used as elements of the hybrid leaping algorithm, and the fitness value corresponding to each training sample is calculated according to the set maximum number of iterations . After comparing the fitness of each training sample, the minimum fitness value is obtained, and this is used as the optimal fitness value.

Among them, the fitness value is a parameter reflecting the pros and cons of the current position of the frog's primary color in the hybrid leapfrog algorithm.

S140. Calculate connection weights by using the optimal fitness values of all road sections, and construct and obtain a traffic flow data analysis model of the wavelet neural network according to the connection weights.

In this step, the connection weight is calculated by using the optimal fitness value obtained in step S130. The connection weight is input into the wavelet neural network to obtain a traffic flow data analysis model of the wavelet neural network. In this embodiment, the traffic flow data analysis model obtained by training the wavelet neural network using the hybrid frog leaping algorithm compares the error with the predicted value and the actual value that only relies on the calculation of the connection weight, and is continuously adjusted according to the magnitude of the error Compared with the practice of wavelet neural network parameters, until the predicted value is getting closer and closer to the true value, it is easier to obtain the connection weight that matches the actual value.

The connection weight includes a first connection weight and a second connection weight. The first connection weight is the connection weight of the input layer of the prediction model constructed by the wavelet neural network, if it represents the connection weight between the i-th node of the input layer and the j-th node of the hidden layer , Can be recorded as _Wij , where j=1, 2,...l, l represents the number of nodes in the hidden layer. The second connection weight is the connection weight of the hidden layer of the prediction model constructed by the wavelet neural network, if it represents the connection weight between the jth node of the hidden layer and the kth node of the output layer The value can be recorded as W _jk , where k=1, 2,...m, and m represents the number of nodes in the output layer.

The method for constructing a traffic flow data analysis model provided by the present application is to obtain traffic flow data on corresponding road sections in a target area through a data channel established with a traffic flow data monitoring system to obtain traffic flow data for different sub-time periods of a set time period. The historical traffic flow data is used as the training sample, and the hybrid leaping algorithm is used to obtain the optimal fitness value of all road sections according to the set maximum number of iterations, and the connection weight of the wavelet neural network is obtained according to the optimal fitness value, Construct a traffic flow data analysis model based on wavelet neural network. This application trains the wavelet neural network through a hybrid leapfrog algorithm to obtain the traffic flow data analysis model. The traffic flow data of the target road section can be obtained by acquiring the traffic flow data of the road section in different periods and different time periods. The prediction solves the problem that the prior art can only simply predict the historical traffic flow in the same time period in the short-term, and cannot meet the long-term traffic flow prediction of the target road section.

For step S130, it may further include the following steps:

A1. Update the training samples of each road segment according to the set maximum number of iterations, and use the hybrid leapfrog algorithm to obtain the updated fitness value of each road segment.

In this step, the maximum number of iterations set by the hybrid leapfrog group algorithm is used to update the training sample x _it of each road section.

According to the updated training sample x _it , the updated fitness value of each road section is calculated.

The specific calculation process of fitness value is as follows:

Fitness value:

Among them, y _i is calculated by the wavelet neural network model, and the formula of the output layer is as follows:

Among them, h(j) represents the output result of the j-th node of the hidden layer, and the selected hidden layer output formula is:

Among them, the fitness value is the prediction error value of the wavelet neural network, y _i is the predicted output of the i-th node, _mi is the expected output of the i-th node, and the expected output is a variable sample of the obtained traffic flow data ; A _i is the expansion factor and b _i is the translation factor.

In order to obtain the fitness value of each road segment, before step S130, that is, before the optimal fitness of the road segment is obtained by the hybrid leaping algorithm, a traffic flow data analysis model of wavelet neural network is required, and the above-mentioned a first wavelet neural network connection weight values W _ij, the second connection weights W _jk, stretching factor a _i, b _i is initialized shift factor is provided, so as to calculate the fitness value of each segment.

A2. According to the updated fitness value, compare the global optimal fitness value and the local fitness value to obtain the optimal fitness value of all the road sections.

According to the updated fitness value, the local optimal fitness value is obtained by comparing all the molecular groups of the fitness values of the road sections, and the global optimal fitness value is obtained by comparing the fitness values of all the road sections . According to the update of the maximum fitness value during each iteration update, the minimum fitness value is calculated with the maximum fitness value obtained by the final update, and the minimum fitness value is used as the optimal fitness value.

For the above-mentioned acquisition and division of the fitness value of the road section, the historical traffic flow data of each road section can be assigned to each road section according to the busy level of the historical traffic flow data in the descending order of the fitness value of each road section in each update iteration. For each busy level, each traffic flow subset is obtained.

Specifically, according to the busy level of historical traffic flow data, it is correspondingly divided into several traffic flow subsets, such as 1, 2, 3...n subsets. And sort according to the fitness value of each road section according to each update iteration from small to large, and get the first 1, 2, 3..., n+m fitness values, and divide the fitness value of the ranking first. Into the first subset, sort the fitness value of the second rank into the second subset,..., sort the fitness value of the nth rank into the nth subset. Finally, each traffic flow subset is obtained.

On this basis, the above step A2 may further include:

A21. Update the maximum fitness value of the current traffic flow subset in all traffic flow subsets according to the maximum number of iterations set by a single traffic flow subset.

For the convenience of calculation, the maximum number of iterations for each traffic flow subset is set to be consistent. According to the maximum number of iterations set by the traffic flow subset, for different sub-time periods within the set time period, the training samples of each road section in each traffic flow subset are updated and iterated, and the corresponding traffic flow subset The current maximum fitness value is updated to obtain the current maximum fitness value in the traffic flow subset.

A22. When the maximum number of iterations set for a single traffic flow subset is less than the number of global hybrid iterations, the global maximum fitness value of the road segment is updated according to the remaining number of iterations.

When updating each traffic flow subset, when the maximum number of iterations set by a single traffic flow subset is met, after all traffic flow subsets have completed the local deep search, if the number of global hybrid iterations is met, update At the end of the process, the last maximum fitness value of the global road section is obtained, and the minimum fitness value is calculated.

The specific calculation process is as follows:

In each iteration, it is first determined that the road section with the largest fitness value of the subset in the current iteration is X _w , the road section with the smallest fitness value is X _b and the road section with the smallest fitness value among all road sections is X _g ; The road section with the largest current fitness value in this subset is X _w for updating. The updating strategy is as follows.

Leapfrog step length update formula:

Ω _i ＝rand().(X _b -X _w )

(|Ω _min ||≤||Ω _i ||≤||Ω _max ||) (1)

The position update formula of the individual frog:

new X _w ＝X _w +Ω _i (2)

Among them, Ω _i represents the update step size of the frog individual, i = 1, 2, ..., N; rand()

It is a random number uniformly distributed between [0,1]; ||Ω _max || represents the maximum leapfrog step allowed to be updated; ||Ω _min || represents the minimum leapfrog step allowed to be updated. Implement the update strategy (1)(2).

If the fitness value of newX _w is less than the fitness value of the original X _w , the updated road segment is used to replace the road segment with the largest fitness in the current subset in the current iteration.

When the local depth search of all traffic flow subsets is completed, all road sections are remixed and sorted and subgroups are divided again, and then the local depth search is performed again, and so on until the number of mixed iterations is met.

If the maximum number of iterations set for a single traffic flow subset is less than the number of global hybrid iterations, then re-sort the latest fitness values obtained from each iteration and then assign them to each traffic flow one by one in the above manner. Concentration, first update the maximum fitness value of all traffic flow subsets. According to the remaining number of iterations, repeat the above calculation and update to obtain the last maximum fitness value for the global road section.

The specific calculation process is as follows:

Leapfrog step length update formula:

Ω _i =rand().(X _g -X _w )

(||Ω _min ||≤||Ω _i ||≤||Ω _max ||) (3)

The position update formula of the individual frog:

new X _w ＝X _w +Ω _i (4)

Implement the update strategy (3)(4).

If the fitness value of newX _w is still not improved, a new X _w is randomly generated.

A23. Calculate the minimum fitness value of the road section according to the updated maximum fitness value, and use the minimum fitness value as the optimal fitness value.

According to the above step A22, the last maximum fitness value is obtained, and the minimum fitness value is calculated and used as the optimal fitness value.

With the help of the hybrid leapfrog algorithm having the characteristics of global search capability, the method for constructing the traffic flow data analysis model provided by this application can converge to the global optimal solution, so it is beneficial to improve the prediction speed and prediction accuracy.

On the premise of the optimal fitness value obtained above, step S140 may further include the following steps:

B1. Using the optimal fitness value of all road sections, obtain the position of the corresponding frog element in the hybrid leapfrog algorithm according to the fitness function curve.

According to the optimal fitness value obtained from step S130 and the curve image of the fitness function of the frog swarm algorithm, the vector value of the green frog position element corresponding to the optimal fitness value is obtained.

In this embodiment, the vector value of the position of the frog element is:

x _id = (W _ij , W _jk ,a _j ,b _j ) (5)

Among them, x _id is the expression form of the vector of x _it .

The fitness function is used to characterize the corresponding relationship between all individuals in the problem and their fitness.

B2. Obtain the optimal solution of the connection weight according to the vector corresponding to the position of the frog element.

The vector of the position of the frog element obtained by using the curve image of the fitness value function is the first connection weight W _ij , the second connection weight W _jk , the expansion factor a _i and the translation factor of the wavelet neural network. The vector of b _i .

The fitness function is used to calculate an optimal solution of the first connection weight W _ij and the second connection weight W _jk according to the optimal fitness value.

The obtained optimal solution of the first connection weight W _ij and the second connection weight W _jk is to be input into the constructed wavelet neural network model to obtain the optimal traffic flow data analysis model for the road section.

The previously obtained test samples of the traffic flow data of each road section are input into the traffic flow data analysis model to obtain the traffic flow prediction data of the target area.

With the help of the hybrid leapfrog algorithm with the characteristics of global search capability, fast convergence speed, and less parameter configuration, the traffic flow data prediction method provided in this application is easier to converge to the global optimal solution, which is beneficial to improve the prediction speed and prediction accuracy.

The sum of the training samples, test samples, and variable samples of traffic flow data involved in the above description is the sum of the traffic flow data acquired in different sub-time periods of the set time period for all road sections, based on empirical values Set the ratio of the above three samples.

In this embodiment, in order to train with as much traffic flow data as possible to obtain the prediction model with better prediction ability, the proportion of the training samples is set to 65%, and the proportion of the variable samples is set to 10%, the proportion of test samples is set to 25%.

The obtained traffic flow data corresponding to each sample is respectively distributed according to proportions and waiting to be entered, and data processing is performed to obtain the traffic flow prediction data of the target area.

Referring to FIG. 2, FIG. 2 is a flowchart of a method for constructing a traffic flow data analysis model according to another embodiment of the application. In order to obtain a more accurate prediction model through training, in this embodiment, in the target area, according to the direction of the traffic volume, all road sections are divided into injecting traffic flow and diverting traffic flow.

Regarding the foregoing division of the traffic flow of the road section, step S130 may further include:

S131: According to the set maximum number of iterations, update the training samples of the traffic flow in each direction of each road section.

Using the maximum number of iterations set by the hybrid leapfrog algorithm, the corresponding training samples of the injected traffic flow and the diversion traffic flow of each road section are updated.

For the actual road segment scenario, the injecting traffic flow and the diversion traffic flow of each road segment are traffic flows in the traffic lanes in opposite directions in a road segment.

S132. Using the hybrid leapfrog algorithm, calculate the corresponding fitness value based on the traffic flow data in each direction of each road section.

According to the updated training samples, the fitness values of each injected traffic flow and diversion traffic flow are calculated respectively.

The specific calculation process refers to the above formulas (1)-(4).

Respectively update and iterate the fitness value of the injected traffic flow and the fitness value of the diversion traffic flow to obtain the optimal fitness value of all the injected traffic flows and the optimal fitness value of all the diversion traffic flows. .

The calculation of the optimal fitness value corresponding to the fitness value of the injected traffic flow or the fitness value of the diversion traffic flow is as follows:

For the division and calculation of the fitness value of the above injecting traffic flow or diverting traffic flow, according to historical traffic flow data, each injecting traffic flow or diverting traffic flow can be divided into busy level, and all injecting traffic flow or diverting traffic can be classified According to the busy level, the flow is divided into different infusion traffic flow sets or diversion traffic flow sets.

C1. According to the maximum number of iterations set for each injected traffic flow set and diversion traffic flow set, update the maximum fitness value of the current injected traffic flow set and diversion traffic flow set in all traffic flow subsets.

C2. When the maximum number of iterations set for the injection traffic flow set and the channeling traffic flow set is less than the global mixed iteration number, the global maximum fitness value of the road segment is updated according to the remaining number of iterations.

C3. Calculate the respective minimum fitness values of the injected traffic flow and the guided traffic flow according to the updated maximum fitness values of the injected traffic flow set and the guided traffic flow set, and use the respective minimum fitness values as The optimal fitness value.

On this basis, the method for constructing the traffic flow data analysis model may include:

D1. Input the traffic flow data in each direction of each road section into the traffic flow data analysis model to obtain traffic flow forecast data of all road sections;

D2. Obtain the traffic flow prediction data of the target area according to the traffic flow prediction data of all road sections.

For all road sections, set a traffic flow direction to be positive, and the traffic flow direction opposite to it to be negative, and input the injected traffic flow and diverted traffic flow of each road section into the traffic flow data analysis model to obtain the traffic in each direction Stream forecast data. According to the direction of the traffic flow, the predicted data of the injected traffic flow and the dredging traffic flow are added together to obtain the predicted data of the traffic flow of all road sections.

According to the prediction data of each road section, the traffic flow prediction data of all road sections in the target area is finally obtained.

In this embodiment, the traffic flow data of the injected traffic flow and the traffic flow data of the diversion traffic flow both include respective corresponding training samples, variable samples, and test samples, and their addition is for the injected traffic flow. And the sum of the traffic flow data obtained in different sub-time periods of the set time period of the diversion traffic flow, and the above three samples of the injection traffic flow and the diversion traffic flow are respectively set in proportions according to empirical values . Moreover, in order to satisfy the corresponding data volume for data processing, the proportion settings of the respective samples corresponding to the injection road section and the dredging road section are the same.

In this embodiment, in order to train with as much traffic flow data as possible to obtain the prediction model with better prediction ability, the proportions of the training samples are all set to 65%, and the proportions of the variable samples are all Set to 10%, and the proportion of test samples is set to 25%.

Based on the same inventive concept as the above-mentioned method for constructing a traffic flow data analysis model, an embodiment of the present application also provides a device for constructing a traffic flow data analysis model, as shown in FIG. 3, including:

The extraction module 310 is configured to establish a data channel with the traffic flow data monitoring system, determine the corresponding road section according to the target area, and extract the traffic flow data of the target area from the data channel;

The training sample setting module 320 is used to obtain historical traffic flow data of each road section in different sub-time periods of the set time period as training samples;

The iterative calculation module 330 is used to update the training samples of each road section according to the set maximum number of iterations, and obtain the optimal fitness value of all road sections by using the hybrid frog leaping algorithm;

The construction module 340 is configured to use the optimal fitness values of all road sections to obtain connection weights, and construct a wavelet neural network traffic flow data analysis model according to the connection weights.

Please refer to FIG. 4, which is a schematic diagram of the internal structure of the server in an embodiment. As shown in FIG. 4, the server includes a processor 410, a storage medium 420, a memory 430, and a network interface 440 connected through a system bus. Wherein, the storage medium 420 of the server stores an operating system, a database, and computer-readable instructions. The database may store control information sequences. When the computer-readable instructions are executed by the processor 410, the processor 410 can realize a traffic In the method for constructing a flow data analysis model, the processor 410 can implement the extraction module 310, the training sample setting module 320, the iterative calculation module 330, and the construction in the device for constructing a traffic flow data analysis model in the embodiment shown in FIG. Function of module 340. The processor 410 of the server is used to provide computing and control capabilities to support the operation of the entire server. The memory 430 of the server may store computer-readable instructions, and when the computer-readable instructions are executed by the processor 410, the processor 410 can execute a method for constructing a traffic flow data analysis model. The network interface 440 of the server is used to connect and communicate with the terminal. Those skilled in the art can understand that the structure shown in FIG. 4 is only a block diagram of part of the structure related to the solution of the present application, and does not constitute a limitation on the server to which the solution of the present application is applied. The specific server may include More or fewer components are shown in the figure, or some components are combined, or have different component arrangements.

In one embodiment, the present application also proposes a storage medium storing computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors perform the following steps: The data channel with the traffic flow data monitoring system determines the corresponding road section according to the target area, and extracts the traffic flow data of the target area from the data channel; obtains the historical traffic flow of each road section in different sub-time periods of the set time period Data is used as a training sample; according to the set maximum number of iterations, the training sample of each road section is updated, and the optimal fitness value of all road sections is obtained using the hybrid frog-leap algorithm; the optimal fitness value of all road sections is used The connection weight is constructed to obtain a traffic flow data analysis model of the wavelet neural network according to the connection weight.

Based on the foregoing embodiments, it can be seen that the greatest beneficial effect of this application lies in:

The technical solution provided by this application uses the hybrid frog leaping algorithm to obtain the connection weights of the wavelet neural network prediction model. It uses the characteristics of the hybrid frog leaping algorithm with strong global search ability, fast convergence speed, and less parameter configuration, making it possible to Fast convergence to obtain the global optimal solution, so the use of the initial wavelet neural network parameters obtained by the hybrid frog leaping algorithm to train the traffic flow data analysis model can achieve the goal of improving the prediction speed and prediction accuracy when predicting the traffic flow. At the same time, the method and device for constructing a traffic flow data analysis model provided by this application have strong global search capabilities and can be well adapted to different traffic flow data samples after updating and iteration, and therefore can be adapted to different traffic flow data samples in different periods and periods. Flow data constructs a corresponding traffic flow data analysis model, filling the gap of long-term traffic flow data analysis tools.

In summary, through the construction and installation of the traffic flow data analysis model, this application will use the hybrid frog leaping algorithm to construct a traffic flow data analysis model of wavelet neural network, which fills in the gap of long-term traffic flow data analysis tools and improves The efficiency of traffic flow data prediction.

A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing relevant hardware through a computer program. The computer program can be stored in a computer readable storage medium. When executed, it may include the processes of the above-mentioned method embodiments. Among them, the aforementioned storage medium may be a storage medium such as a magnetic disk, an optical disc, a read-only memory (Read-Only Memory, ROM), or a random access memory (Random Access Memory, RAM), etc.

The technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above-mentioned embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, All should be considered as the scope of this specification.

The above-mentioned embodiments only express several implementation manners of this application, and their descriptions are more specific and detailed, but they should not be construed as limiting the scope of this application. It should be pointed out that for those of ordinary skill in the art, without departing from the concept of this application, several modifications and improvements can be made, and these all fall within the protection scope of this application. Therefore, the scope of protection of the patent of this application shall be subject to the appended claims.

Claims

A method for constructing a traffic flow data analysis model is characterized in that it comprises the following steps:

Establish a data channel with the traffic flow data monitoring system, determine the corresponding road section according to the target area, and extract the traffic flow data of the target area from the data channel;

Obtain historical traffic flow data of different sub-time periods of each road section in the set time period as training samples;

According to the set maximum number of iterations, the training samples of each road section are updated, and the hybrid frog leaping algorithm is used to obtain the optimal fitness value of all road sections;

Use the optimal fitness values of all road sections to obtain connection weights, and construct a wavelet neural network traffic flow data analysis model according to the connection weights;

The traffic flow data of each road section corresponding to the target area is input into the traffic flow data analysis model to obtain traffic flow prediction data of the target area.
The method according to claim 1, wherein:

The step of updating the training samples of each road section according to the set maximum number of iterations, and obtaining the optimal fitness value of all road sections by using the hybrid frog leaping algorithm includes:

According to the set maximum number of iterations, update the training samples of each road segment, and use the hybrid leapfrog algorithm to obtain the updated fitness value of each road segment;

According to the updated fitness value, it is compared with the global optimal fitness value and the local fitness value to obtain the optimal fitness values of all the road sections.
The method according to claim 2, wherein:

According to the busy level of the historical traffic flow data of each said road segment, the fitness value of each said road segment in each update iteration is allocated to each busy level in ascending order to obtain each traffic flow subset;

The step of comparing the global optimal fitness value and the local fitness value according to the updated fitness value to obtain the optimal fitness value of all road sections includes:

Update the maximum fitness value of the current traffic flow subset in all traffic flow subsets according to the maximum number of iterations set by a single traffic flow subset;

When the maximum number of iterations set for a single traffic flow subset is less than the number of global hybrid iterations, the global maximum fitness value of the road segment is updated according to the remaining number of iterations;

According to the updated maximum fitness value, the minimum fitness value of the road section is calculated, and the minimum fitness value is used as the optimal fitness value.
The method according to claim 3, characterized in that:

The step of using the optimal fitness values of all road sections to obtain connection weights, and constructing and obtaining a traffic flow data analysis model according to the connection weights includes:

Use the optimal fitness value of all road sections to obtain the position of the corresponding frog element in the hybrid leapfrog algorithm according to the fitness function curve;

Obtaining the optimal solution of the connection weight according to the vector corresponding to the position of the frog element;

According to the optimal solution of the connection weight, a traffic flow data analysis model is constructed.
The method according to claim 1, wherein:

In the target area, according to the direction of the traffic volume, divide all road sections into injecting traffic flow and diverting traffic flow;

The steps of updating the training samples of each road section according to the set maximum number of iterations, and obtaining the optimal fitness value of all road sections by using the hybrid leapfrog algorithm, include:

According to the set maximum number of iterations, update the training samples of the traffic flow in each direction of each road section;

Using the hybrid leapfrog algorithm, the corresponding fitness value is calculated based on the traffic flow data of each road section in each direction.
The method according to claim 5, wherein the traffic flow data of each road section corresponding to the target area is input into the traffic flow data analysis model to obtain the traffic flow data of the target area Forecast data, including:

Input the traffic flow data in each direction of each road section into the traffic flow data analysis model to obtain traffic flow forecast data of all road sections;

According to the traffic flow prediction data of all road sections, the traffic flow prediction data of the target area is obtained.
The method according to claim 1, wherein:

Before the step of updating the training samples of each road segment according to the set maximum number of iterations, and obtaining the optimal fitness value of all road segments by using the hybrid leaping algorithm, it also includes:

For the traffic flow data analysis model, the connection weights, expansion factors, and translation factors are initialized according to experience values.
A device for constructing a traffic flow data analysis model is characterized in that it comprises:

The extraction module is used to establish a data channel with the traffic flow data monitoring system, determine the corresponding road section according to the target area, and extract the traffic flow data of the target area from the data channel;

The training sample setting module is used to obtain the historical traffic flow data of each road section in different sub-time periods of the set time period as training samples;

The iterative calculation module is used to update the training samples of each road section according to the set maximum number of iterations, and obtain the optimal fitness value of all road sections by using the hybrid leapfrog algorithm;

A construction module for obtaining connection weights using the optimal fitness values of all road sections, and constructing and obtaining a wavelet neural network traffic flow data analysis model according to the connection weights;

The building module is also used to input the traffic flow data of each road section corresponding to the target area into the traffic flow data analysis model to obtain traffic flow prediction data in the target area.
The device according to claim 8, wherein the iterative calculation module is specifically configured to:

According to the set maximum number of iterations, update the training samples of each road segment, and use the hybrid leapfrog algorithm to obtain the updated fitness value of each road segment;

According to the updated fitness value, it is compared with the global optimal fitness value and the local fitness value to obtain the optimal fitness values of all the road sections.
The device according to claim 9, wherein:

The iterative calculation module is further configured to assign each busy level to each busy level according to the busy level of the historical traffic flow data of each of the road sections, according to the fitness value of each road section in each update iteration from small to large, Get a subset of each traffic flow;

When the iterative calculation module compares the global optimal fitness value and the local fitness value according to the updated fitness value to obtain the optimal fitness value of all road sections, it is specifically used for:

Update the maximum fitness value of the current traffic flow subset in all traffic flow subsets according to the maximum number of iterations set by a single traffic flow subset;

When the maximum number of iterations set for a single traffic flow subset is less than the number of global hybrid iterations, the global maximum fitness value of the road segment is updated according to the remaining number of iterations;

According to the updated maximum fitness value, the minimum fitness value of the road section is calculated, and the minimum fitness value is taken as the optimal fitness value.
The device according to claim 10, wherein the building module is specifically configured to:

Use the optimal fitness value of all road sections to obtain the position of the corresponding frog element in the hybrid leapfrog algorithm according to the fitness function curve;

Obtaining the optimal solution of the connection weight according to the vector corresponding to the position of the frog element;

According to the optimal solution of the connection weight, a traffic flow data analysis model is constructed.
The device according to claim 8, wherein:

The training sample setting module is also used to divide all road sections into injecting traffic flow and diverting traffic flow in the target area according to the direction of traffic volume;

The iterative calculation module is specifically used for:

According to the set maximum number of iterations, update the training samples of the traffic flow in each direction of each road section;

Using the hybrid leapfrog algorithm, the corresponding fitness value is calculated based on the traffic flow data of each road section in each direction.
The device according to claim 12, wherein:

The building module is also used to input the traffic flow data in each direction of each road section into the traffic flow data analysis model to obtain traffic flow prediction data for all road sections; and obtain traffic flow prediction data for all road sections Forecast data of traffic flow in the target area.
The device according to claim 8, wherein:

The iterative calculation module is also used to update the training samples of each road section according to the set maximum number of iterations, and use the hybrid frog leaping algorithm to obtain the optimal fitness value of all road sections, for traffic flow data Analyze the model, and initialize the connection weight, expansion factor, and translation factor according to experience values.
A server, characterized in that it comprises:

One or more processors;

Memory

One or more computer programs, wherein the one or more computer programs are stored in the memory and configured to be executed by the one or more processors, and the one or more computer programs are configured to execute The following steps:

Establish a data channel with the traffic flow data monitoring system, determine the corresponding road section according to the target area, and extract the traffic flow data of the target area from the data channel;

Obtain historical traffic flow data of different sub-time periods of each road section in the set time period as training samples;

According to the set maximum number of iterations, the training samples of each road section are updated, and the hybrid frog leaping algorithm is used to obtain the optimal fitness value of all road sections;

Use the optimal fitness values of all road sections to obtain connection weights, and construct a wavelet neural network traffic flow data analysis model according to the connection weights;

The traffic flow data of each road section corresponding to the target area is input into the traffic flow data analysis model to obtain traffic flow prediction data of the target area.
The server according to claim 15, wherein the training samples of each road section are updated according to the set maximum number of iterations, and the hybrid leapfrog algorithm is used to obtain the optimal fitness value of all road sections. , The one or more computer programs are configured to perform the following steps:

According to the set maximum number of iterations, update the training samples of each road segment, and use the hybrid leapfrog algorithm to obtain the updated fitness value of each road segment;

According to the updated fitness value, it is compared with the global optimal fitness value and the local fitness value to obtain the optimal fitness values of all the road sections.
The server of claim 16, wherein the one or more computer programs are further configured to perform the following steps:

According to the busy level of the historical traffic flow data of each said road segment, the fitness value of each said road segment in each update iteration is allocated to each busy level in ascending order to obtain each traffic flow subset;

When the updated fitness value is compared with the global optimal fitness value and the local fitness value to obtain the optimal fitness values of all road sections, the one or more computer programs are configured to Perform the following steps:

Update the maximum fitness value of the current traffic flow subset in all traffic flow subsets according to the maximum number of iterations set by a single traffic flow subset;

When the maximum number of iterations set for a single traffic flow subset is less than the number of global hybrid iterations, the global maximum fitness value of the road segment is updated according to the remaining number of iterations;

According to the updated maximum fitness value, the minimum fitness value of the road section is calculated, and the minimum fitness value is used as the optimal fitness value.
The server according to claim 15, wherein the one or more computer programs are further configured to perform the following steps:

In the target area, according to the direction of traffic volume, divide all road sections into injecting traffic flow and diverting traffic flow;

When the training samples of each road section are updated according to the set maximum number of iterations, and the hybrid leaping algorithm is used to obtain the optimal fitness value of all road sections, the one or more computer programs are configured to execute the following step:

According to the set maximum number of iterations, update the training samples of the traffic flow in each direction of each road section;

Using the hybrid leapfrog algorithm, the corresponding fitness value is calculated based on the traffic flow data of each road section in each direction.
The server according to claim 18, wherein the one or more computer programs are further configured to perform the following steps:

Input the traffic flow data in each direction of each road section into the traffic flow data analysis model to obtain traffic flow forecast data of all road sections;

According to the traffic flow prediction data of all road sections, the traffic flow prediction data of the target area is obtained.
A computer non-volatile readable storage medium, characterized in that a computer program is stored on the computer non-volatile readable storage medium, and the computer program implements any one of claims 1-7 when executed by a processor The method for constructing the traffic flow data analysis model.