Road section data screening method, system and storage medium

Abstract

The present invention provides a road section data screening method, a road section data screening system and a computer-readable storage medium. The method includes: obtaining trajectory data of historical orders; mining the trajectory data using a sequence mining algorithm to determine multiple road section data; determining the yaw rate of the historical order under any road section data among the multiple road section data; screening the road section data according to the yaw rate. By using the technical solution of the present application, compared with the method of determining key road section data only by traffic in the related technical solution, there is no need to process the road section data screened according to traffic one by one, which reduces the amount of data to be processed, thereby reducing the demand for computing power, and can select processing hardware with lower computing power to implement hardware support. At the same time, since the amount of data to be processed is reduced, energy consumption can be effectively reduced.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a road section data screening method, a road section data screening system and a computer-readable storage medium. Background Art

[0002] Generally speaking, in the relevant technologies, the research on the operation status of the road network is usually based on the changes in the road network flow, such as the research on the routes in the transportation field and the network community field. However, the method of determining the road section data based on the road network flow requires processing a large amount of data, and cannot accurately determine the road section data that affects the vehicle deviation in the road network, and thus cannot meet the current analysis needs. Summary of the invention

[0003] The present invention aims to solve at least one of the technical problems existing in the prior art or related art.

[0004] To this end, in an embodiment of the first aspect of the present invention, a method for screening road section data is proposed, including: obtaining trajectory data of historical orders; mining the trajectory data using a sequence mining algorithm to determine multiple road section data; determining the yaw rate of the historical order under any road section data among the multiple road section data; and screening the road section data according to the yaw rate.

[0005] In this technical solution, the trajectory data of historical orders are obtained, wherein the historical orders are historical travel data of online car-hailing, rental vehicles, private vehicles, etc., and the trajectory data are mined by using a sequence mining algorithm to determine multiple road segment data, and multiple road segments with a certain support are selected from the trajectory data, wherein each of the multiple road segments has a key OD (Origin Destination, departure place and travel destination), but it should be noted that OD is only the departure place and travel destination of each section in the trajectory data of the historical order, rather than the departure place and destination of the entire historical order route, determine the yaw rate of the historical order under each section data in the multiple section data, and filter the section data from the multiple section data according to the yaw rate, wherein the yaw rate is the ratio of the number of yaw orders in the historical orders passing through the section data to the historical orders, use the yaw rate to filter the multiple section data, and filter out the key section data that can reflect the yaw problem, thereby avoiding the inaccuracy of selecting key sections for yaw problem research only through road traffic, and obtain the section data through screening for research, thereby improving the accuracy of the research results for the section, wherein, by using the technical solution of the present application, compared with the method of determining key section data only through traffic in the related technical solution, there is no need to process the section data filtered according to traffic one by one, thereby reducing the amount of data to be processed, thereby reducing the demand for computing power, and can select processing hardware with lower computing power to implement hardware support, and at the same time, due to the reduction in processing volume, it can effectively reduce energy consumption.

[0006] In addition, the road section data screening method according to the above embodiment of the present invention may also have the following additional technical features:

[0007] In the above technical solution, further, the step of filtering the road section data according to the yaw rate specifically includes: sorting the multiple road section data according to the yaw rate, and selecting the road section data corresponding to the maximum yaw rate as the first road section data.

[0008] In this technical solution, multiple sections are screened by yaw rate. Therefore, the yaw rate is used to sort multiple sections during the screening process. Since the first section data is the section data corresponding to the largest yaw rate, it can better reflect the cause of yaw than other section data in the process of studying the yaw problem, so it can be judged whether the road is abnormal based on the first section data. Selecting the section data with the largest yaw rate as the first section data avoids analyzing the cause of yaw for multiple section data, reduces the amount of data that needs to be processed for yaw cause analysis, and improves the speed of determining abnormal section data while reducing computing power occupation and energy consumption.

[0009] In any of the above technical solutions, it further includes: classifying historical orders according to the first road section data; when the yaw rate of historical orders whose trajectory data does not contain the first road section data is greater than the first yaw value, or when the number of historical orders whose trajectory data does not contain the first road section data is greater than the first order value: selecting one road section data from multiple road section data that do not contain the first road section data as the second road section data.

[0010] In this technical solution, after determining the first section data, the first section data is used to classify historical orders to obtain historical orders containing the first section data and historical orders not containing the first section data. In order to eliminate the screening of later section data by the first section data, historical orders not containing the first section data are selected to conduct a new round of screening of section data. Prior to this, it is determined whether the historical orders not containing the first section data meet the conditions for screening the section data again. Specifically, when the yaw rate of historical orders whose trajectory data do not contain the first section data is less than the first yaw value, that is, the probability of yaw in historical orders not containing the first section data is low, in order to reduce the occupation of computing power and energy consumption. , no new round of screening of road section data is performed. Similarly, when the number of historical orders in which the trajectory data does not contain the first road section data is less than the first order value, the number of historical orders in which the trajectory data does not contain the first road section data is too small. In order to reduce the occupation of computing power and energy consumption, no new round of screening of road section data is performed. Only when the yaw rate of historical orders in which the trajectory data does not contain the first road section data is greater than the first yaw value, or when the number of historical orders in which the trajectory data does not contain the first road section data is greater than the first order value, one road section data is selected from multiple road section data that do not contain the first road section data as the second road section data, and then the road section data reflecting the cause of the yaw is screened for analysis. The above steps of the technical solution can eliminate the influence of the first road section data on the screening of the second road section data, and then the second road section data obtained can truly reflect the cause of the yaw, which is convenient for finding abnormal sections.

[0011] In any of the above technical solutions, further, after the step of selecting any one section of road data from multiple section data that does not contain the first section data as the second section data, it also includes: taking the historical orders that do not contain the first section data as the updated historical orders; and classifying the updated historical orders according to the second section data until the yaw rate of the updated historical orders is less than the first yaw value or the number of updated historical orders is less than the first order value.

[0012] In this technical solution, by taking the historical orders that do not contain the first road section data as the updated historical orders, the previous judgment criteria can be used in the new round of judgment process. Since the judgment conditions for judging whether a new round of road section data screening is needed are the same, the road section data screening process can be performed multiple times until the conditions for further screening are not met. The above process is classified using the idea of ​​decision trees. Through execution, road section data that do not affect each other can be screened out to analyze the cause of deviation, avoiding the problem of needing to repeatedly analyze the same historical order multiple times at this stage, which requires a large amount of data processing, and at the same time avoiding the road section data screened out being associated with each other, affecting the analysis of the cause of deviation.

[0013] In any of the above technical solutions, further, the step of selecting one road section data from multiple road section data that do not include the first road section data as the second road section data specifically includes: sorting the multiple road section data that do not include the first road section data based on the yaw rate to obtain a sorting result, and selecting a preset number of road section data according to the sorting result; and selecting one road section data from the preset number of road section data as the second road section data.

[0014] In this technical solution, in order to improve the screening speed of the road section data, the number of multiple road section data is limited in the process of a new round of road section data screening, further, the multiple road section data that do not include the first road section data are sorted based on the yaw rate, and a preset number of road section data are selected as candidate road section data according to the sorting result, and the second road section data are screened out based on the candidate road section data. Further, the preset number is the result of multiple experiments, and the candidate road section data obtained by screening is the road section data obtained after selecting the preset number from high to low in the sorting result.

[0015] In any of the above technical solutions, further, it also includes: when it is confirmed that an unordered set of road section data obtained by filtering according to the yaw rate exists, stopping the classification of historical orders.

[0016] In this technical solution, in order to further reduce the time spent on the screening process of the road section data, it is determined whether an unordered set composed of the road section data screened according to the yaw rate exists. If the determined unordered set exists, the classification of the historical orders is stopped. By determining whether the unordered set exists, if the unordered sets corresponding to the determined multiple road section data already exist, the classification of the historical orders is stopped in advance, thereby avoiding the screening of the road section data, reducing the amount of data to be processed and the energy consumption.

[0017] In any of the above technical solutions, further, the step of confirming the existence of the unordered set composed of the selected road section data specifically includes: searching for the unique code corresponding to the road section data obtained by filtering according to the yaw rate; if the unordered set composed of the unique code exists, it is determined that the unordered set composed of the road section data obtained by filtering according to the yaw rate exists.

[0018] In this technical solution, by searching for the unique code of the road section data, using the code to construct an unordered set, and determining whether the unordered set formed by the unique code exists. Specifically, the method of determining the unordered set formed by the unique code is simple, and the computing power required is low compared to the unordered set formed by directly using the selected road section data, and the amount of data processed is small, so that less time is required to obtain the same determination result, thereby improving the determination speed.

[0019] In any of the above technical solutions, it further includes: when the ratio of the number of historical orders whose trajectory data does not contain the road section data obtained by filtering the yaw rate to the number of historical orders is less than a third preset threshold, the classification of historical orders is stopped; or when the number of historical orders whose trajectory data does not contain the road section data obtained by filtering the yaw rate is less than a fourth preset threshold, the classification of historical orders is stopped.

[0020] In this technical solution, in order to further improve the screening speed of road section data, it is necessary to propose new pruning conditions for classifying historical orders. For example, when the ratio of the number of historical orders whose trajectory data does not contain road section data obtained by filtering according to the yaw rate to the number of historical orders is less than a third preset threshold, the classification of historical orders is stopped; or when the number of historical orders whose trajectory data does not contain road section data obtained by filtering according to the yaw rate is less than a fourth preset threshold, the classification of historical orders is stopped, thereby reducing the demand for computing power and energy consumption in the process of classifying historical orders.

[0021] In any of the above technical solutions, further, the trajectory data includes actual walking trajectories and navigation trajectories, and the trajectory data is mined using a sequence mining algorithm to determine the steps of multiple road section data, specifically including: using a sequence mining algorithm to mine the actual walking trajectory and the navigation trajectory to determine a first group of road section data sets and a second group of road section data sets; determining multiple road section data based on the intersection of the first group of road section data sets and the second group of road section data sets.

[0022] In this technical solution, the trajectory data is the historical travel data of online car-hailing, rental vehicles, private vehicles, etc., wherein the trajectory data includes two situations, the trajectory data is a navigation trajectory or the trajectory data is an actual travel trajectory. Therefore, when determining multiple road section data, firstly determine the first group of road section data sets according to whether the trajectory data is an actual travel trajectory, determine the second group of road section data sets according to whether the trajectory data is a navigation trajectory, and determine multiple road section data by the intersection of the first group of road section data sets and the second group of road section data sets; or first determine the second group of road section data sets according to whether the trajectory data is an actual travel trajectory, determine the first group of road section data sets according to whether the trajectory data is a navigation trajectory, and determine multiple road section data by the intersection of the first group of road section data sets and the second group of road section data sets, and filter the road section data according to the historical orders with deviation, and then analyze the reasons for the deviation.

[0023] In any of the above technical solutions, further, the support degree of the first group of road section data and the second group of road section data is greater than a preset threshold; or the support degree of multiple road section data is greater than a preset threshold.

[0024] In this technical solution, in order to further improve the screening speed of road section data, while determining multiple road section data, the support of the sequence mining algorithm is set to filter out the road section data with a support less than a preset threshold, and only the road section data with a support greater than the preset threshold is screened, thereby reducing the number of multiple road section data, reducing the amount of data processing, and reducing energy consumption.

[0025] In any of the above technical solutions, further, the sequence mining algorithm is a binary PrefixSpan algorithm.

[0026] In this technical solution, PrefixSpan (Prefix-Projected Pattern Growth, i.e., a prefix projection mining algorithm) may also select other data mining algorithms during specific use.

[0027] In an embodiment of the second aspect of the present invention, a road section data screening system is proposed, comprising: a memory configured to store executable instructions; a processor configured to execute the stored instructions to implement the steps of any of the above-mentioned road section data screening methods.

[0028] Specifically, the processor is configured to execute executable instructions stored in the memory to: obtain trajectory data of historical orders; use a sequence mining algorithm to mine the trajectory data to determine multiple road segment data; determine the yaw rate of the historical order under any road segment data among the multiple road segment data; and filter the road segment data according to the yaw rate.

[0029] In the technical solution, the processor is configured to execute executable instructions stored in the memory to: obtain trajectory data of historical orders, wherein the historical orders are historical travel data of online car-hailing, rental vehicles, private vehicles, etc., mine the trajectory data by using a sequence mining algorithm, determine multiple road segment data, select multiple road segments with a certain support from the trajectory data, wherein each of the multiple road segments has a key OD (Origin Destination, departure place and travel destination), but it should be noted that OD is only the departure place and travel destination of each section in the trajectory data of the historical order, rather than the departure place and destination of the entire historical order route, and the yaw rate of the historical order under each section data in the multiple section data is determined, and the section data is obtained by screening from the multiple section data according to the yaw rate, wherein the yaw rate is the ratio of the number of yaw orders in the historical orders passing through the section data to the historical orders, and the multiple section data are screened using the yaw rate to screen out the key section data that can reflect the yaw problem, thereby avoiding the inaccuracy of selecting key sections for yaw problem research only through road traffic, and the section data is obtained through screening for research, thereby improving the accuracy of the research results for the sections, wherein, by using the technical solution of the present application, compared with the method of determining key section data only through traffic in the related technical solution, there is no need to process the section data screened according to traffic one by one, thereby reducing the amount of data to be processed, thereby reducing the demand for computing power, and can select processing hardware with lower computing power to achieve hardware support, and at the same time, due to the reduction in processing volume, it can effectively reduce energy consumption.

[0030] In addition, the road section data screening system according to the above embodiment of the present invention may also have the following additional technical features:

[0031] In the above technical solution, further, the processor is configured to execute executable instructions stored in the memory to: sort the multiple road section data according to the yaw rate, and select the road section data corresponding to the maximum yaw rate as the first road section data.

[0032] In the technical solution, the processor is configured to execute the executable instructions stored in the memory to: screen multiple road sections by yaw rate, so that the yaw rate is used to sort the multiple road sections during the screening process. Since the first road section data is the road section data corresponding to the largest yaw rate, it can better reflect the yaw cause than other road section data in the process of studying the yaw problem, so it can be judged whether the road is abnormal based on the first road section data. The road section data with the largest yaw rate is selected as the first road section data, which avoids analyzing the yaw cause for multiple road section data, reduces the amount of data to be processed for the yaw cause analysis, and improves the speed of determining abnormal road section data while reducing the computing power occupation and energy consumption.

[0033] In any of the above technical solutions, further, the processor is also configured to execute executable instructions stored in the memory to: classify historical orders according to the first road section data; when the yaw rate of historical orders whose trajectory data does not contain the first road section data is greater than the first yaw value, or when the number of historical orders whose trajectory data does not contain the first road section data is greater than the first order value: select one road section data from multiple road section data that do not contain the first road section data as the second road section data.

[0034] In this technical solution, after determining the first section data, the first section data is used to classify historical orders to obtain historical orders containing the first section data and historical orders not containing the first section data. In order to eliminate the screening of later section data by the first section data, historical orders not containing the first section data are selected to conduct a new round of screening of section data. Prior to this, it is determined whether the historical orders not containing the first section data meet the conditions for screening the section data again. Specifically, when the yaw rate of historical orders whose trajectory data do not contain the first section data is less than the first yaw value, that is, the probability of yaw in historical orders not containing the first section data is low, in order to reduce the occupation of computing power and energy consumption. , no new round of screening of road section data is performed. Similarly, when the number of historical orders in which the trajectory data does not contain the first road section data is less than the first order value, the number of historical orders in which the trajectory data does not contain the first road section data is too small. In order to reduce the occupation of computing power and energy consumption, no new round of screening of road section data is performed. Only when the yaw rate of historical orders in which the trajectory data does not contain the first road section data is greater than the first yaw value, or when the number of historical orders in which the trajectory data does not contain the first road section data is greater than the first order value, one road section data is selected from multiple road section data that do not contain the first road section data as the second road section data, and then the road section data reflecting the cause of the yaw is screened for analysis. The above steps of the technical solution can eliminate the influence of the first road section data on the screening of the second road section data, and then the second road section data obtained can truly reflect the cause of the yaw, which is convenient for finding abnormal sections.

[0035] In any of the above technical solutions, further, the processor is also configured to execute executable instructions stored in the memory to: take historical orders that do not contain the first road section data as updated historical orders; classify the updated historical orders according to the second road section data until the yaw rate of the updated historical orders is less than the first yaw value or the number of updated historical orders is less than the first order value.

[0036] In this technical solution, by taking the historical orders that do not contain the first road section data as the updated historical orders, the previous judgment criteria can be used in the new round of judgment process. Since the judgment conditions for judging whether a new round of road section data screening is needed are the same, the road section data screening process can be performed multiple times until the conditions for further screening are not met. The above process is classified using the idea of ​​decision trees. Through execution, road section data that do not affect each other can be screened out to analyze the cause of deviation, avoiding the problem of needing to repeatedly analyze the same historical order multiple times at this stage, which requires a large amount of data processing, and at the same time avoiding the road section data screened out being associated with each other, affecting the analysis of the cause of deviation.

[0037] In any of the above technical solutions, further, the processor is configured to execute executable instructions stored in the memory to: sort multiple road section data that do not include the first road section data based on the yaw rate to obtain a sorting result, and select a preset number of road section data according to the sorting result; and select any one road section data from the preset number of road section data as the second road section data.

[0038] In this technical solution, in order to improve the screening speed of the road section data, the number of multiple road section data is limited in the process of a new round of road section data screening, further, the multiple road section data that do not include the first road section data are sorted based on the yaw rate, and a preset number of road section data are selected as candidate road section data according to the sorting result, and the second road section data are screened out based on the candidate road section data. Further, the preset number is the result of multiple experiments, and the candidate road section data obtained by screening is the road section data obtained after selecting the preset number from high to low in the sorting result.

[0039] In any of the above technical solutions, further, the processor is also configured to execute executable instructions stored in the memory to: stop classifying historical orders when it is confirmed that an unordered set of road segment data obtained by filtering according to the yaw rate exists.

[0040] In this technical solution, in order to further reduce the time spent on the screening process of the road section data, it is determined whether an unordered set composed of the road section data screened according to the yaw rate exists. If the determined unordered set exists, the classification of the historical orders is stopped. By determining whether the unordered set exists, if the unordered sets corresponding to the determined multiple road section data already exist, the classification of the historical orders is stopped in advance, thereby avoiding the screening of the road section data, reducing the amount of data to be processed and the energy consumption.

[0041] In any of the above technical solutions, further, the processor is configured to execute executable instructions stored in the memory to: search for a unique code corresponding to the road section data obtained by filtering according to the yaw rate; if an unordered set composed of unique codes exists, it is determined that an unordered set composed of road section data obtained by filtering according to the yaw rate exists.

[0042] In this technical solution, by searching for the unique code of the road section data, using the code to construct an unordered set, and determining whether the unordered set formed by the unique code exists. Specifically, the method of determining the unordered set formed by the unique code is simple, and the computing power required is low compared to the unordered set formed by directly using the selected road section data, and the amount of data processed is small, so that less time is required to obtain the same determination result, thereby improving the determination speed.

[0043] In any of the above technical solutions, further, the processor is also configured to execute executable instructions stored in the memory to: stop classifying historical orders when the ratio of the number of historical orders whose trajectory data does not contain road segment data obtained by filtering according to the yaw rate to the number of historical orders is less than a third preset threshold, or stop classifying historical orders when the number of historical orders whose trajectory data does not contain road segment data obtained by filtering according to the yaw rate is less than a fourth preset threshold.

[0044] In this technical solution, in order to further improve the screening speed of road section data, it is necessary to propose new pruning conditions for classifying historical orders. For example, when the ratio of the number of historical orders whose trajectory data does not contain road section data obtained by filtering according to the yaw rate to the number of historical orders is less than a third preset threshold, the classification of historical orders is stopped; or when the number of historical orders whose trajectory data does not contain road section data obtained by filtering according to the yaw rate is less than a fourth preset threshold, the classification of historical orders is stopped, thereby reducing the demand for computing power and energy consumption in the process of classifying historical orders.

[0045] In any of the above technical solutions, further, the trajectory data includes actual walking trajectories and navigation trajectories, and the processor is configured to execute executable instructions stored in the memory to: use a sequence mining algorithm to mine the actual walking trajectories and the navigation trajectories to determine a first set of road segment data sets and a second set of road segment data sets; determine multiple road segment data based on the intersection of the first set of road segment data sets and the second set of road segment data sets.

[0046] In this technical solution, the trajectory data is the historical travel data of online car-hailing, rental vehicles, private vehicles, etc., wherein the trajectory data includes two situations, the trajectory data is a navigation trajectory or the trajectory data is an actual travel trajectory. Therefore, when determining multiple road section data, firstly determine the first group of road section data sets according to whether the trajectory data is an actual travel trajectory, determine the second group of road section data sets according to whether the trajectory data is a navigation trajectory, and determine multiple road section data by the intersection of the first group of road section data sets and the second group of road section data sets; or first determine the second group of road section data sets according to whether the trajectory data is an actual travel trajectory, determine the first group of road section data sets according to whether the trajectory data is a navigation trajectory, and determine multiple road section data by the intersection of the first group of road section data sets and the second group of road section data sets, and filter the road section data according to the historical orders with deviation, and then analyze the reasons for the deviation.

[0047] In any of the above technical solutions, further, the support degree of the first group of road section data and the second group of road section data is greater than a preset threshold; or the support degree of multiple road section data is greater than a preset threshold.

[0048] In this technical solution, in order to further improve the screening speed of road section data, while determining multiple road section data, the support of the sequence mining algorithm is set to filter out the road section data with a support less than a preset threshold, and only the road section data with a support greater than the preset threshold is screened, thereby reducing the number of multiple road section data, reducing the amount of data processing, and reducing energy consumption.

[0049] In any of the above technical solutions, further, the sequence mining algorithm is a binary PrefixSpan algorithm.

[0050] In this technical solution, PrefixSpan (Prefix-Projected Pattern Growth, i.e., a prefix projection mining algorithm) may also select other data mining algorithms during specific use.

[0051] In an embodiment of the third aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored, and the computer program is executed by a processor to implement any of the above-mentioned road section data screening methods.

[0052] The present invention provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above-mentioned road section data screening methods are implemented. Therefore, it has all the technical effects of any of the above-mentioned road section data screening methods, which will not be repeated here.

[0053] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0055] Figure 1 A schematic diagram showing a flow chart of a method for screening road segment data according to an embodiment of the present invention;

[0056] Figure 2 A schematic diagram showing a flow chart of a method for screening road segment data according to another embodiment of the present invention;

[0057] Figure 3 A schematic diagram showing a flow chart of a method for screening road segment data according to yet another embodiment of the present invention;

[0058] Figure 4 A schematic diagram showing a flow chart of a method for screening road segment data according to yet another embodiment of the present invention;

[0059] Figure 5 A schematic diagram of a process of selecting one road section data from a plurality of road section data not including the first road section data as the second road section data is shown according to an embodiment of the present invention;

[0060] Figure 6 A schematic diagram showing a flow chart of a method for screening road segment data according to yet another embodiment of the present invention;

[0061] Figure 7 A schematic diagram of a process for confirming the existence of an unordered set of selected road segment data components according to an embodiment of the present invention is shown;

[0062] Figure 8 A schematic diagram showing a flow chart of a method for screening road segment data according to yet another embodiment of the present invention;

[0063] Fig. 9 A schematic diagram of a process of mining trajectory data using a sequence mining algorithm to determine multiple road segment data in one embodiment of the present invention is shown;

[0064] Fig.10 A schematic diagram of the structure of a decision tree according to an embodiment of the present invention is shown;

[0065] Fig.11 A schematic block diagram of a road segment data screening system according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0066] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0067] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0068] In an embodiment of the first aspect of the present invention, Figure 1 As shown, the road section data screening method includes:

[0069] Step S102, obtaining the trajectory data of historical orders;

[0070] Step S104, using a sequence mining algorithm to mine the trajectory data to determine multiple road section data;

[0071] Step S106, determining the yaw rate of a historical order under any road section data among the plurality of road section data;

[0072] Step S108, filtering the road section data according to the yaw rate.

[0073] In this embodiment, the trajectory data of historical orders are obtained, wherein the historical orders are historical travel data of online car-hailing, rental vehicles, private vehicles, etc., and the trajectory data is mined by using a sequence mining algorithm to determine multiple road segment data, and multiple road segments with a certain support are selected from the trajectory data, wherein each of the multiple road segments has a key OD (Origin Destination, departure place and travel destination), but it should be noted that OD is only the departure place and travel destination of each section in the trajectory data of the historical order, rather than the departure place and destination of the entire historical order route, determine the yaw rate of the historical order under each section data in the multiple section data, and filter the section data from the multiple section data according to the yaw rate, wherein the yaw rate is the ratio of the number of yaw orders in the historical orders passing through the section data to the historical orders, use the yaw rate to filter the multiple section data, and filter out the key section data that can reflect the yaw problem, thereby avoiding the inaccuracy of selecting key sections for yaw problem research only through road traffic, and obtain the section data through screening for research, thereby improving the accuracy of the research results for the section, wherein, by using the technical solution of the present application, compared with the method of determining key section data only through traffic in the related technical solution, there is no need to process the section data filtered according to traffic one by one, thereby reducing the amount of data to be processed, thereby reducing the demand for computing power, and can select processing hardware with lower computing power to implement hardware support, and at the same time, due to the reduction in processing volume, it can effectively reduce energy consumption.

[0074] In one embodiment of the present invention, Figure 2 As shown, the road section data screening method includes:

[0075] Step S202, obtaining the trajectory data of historical orders;

[0076] Step S204, using a sequence mining algorithm to mine the trajectory data to determine multiple road section data;

[0077] Step S206, determining the yaw rate of a historical order under any road section data among the plurality of road section data;

[0078] Step S208 , sorting the plurality of road section data according to the yaw rate, and selecting the road section data corresponding to the maximum yaw rate as the first road section data.

[0079] In this embodiment, multiple sections are screened by yaw rate. Therefore, the yaw rate is used to sort multiple sections in the screening process. Since the first section data is the section data corresponding to the largest yaw rate, it can better reflect the cause of yaw than other section data in the process of studying the yaw problem, so it can be judged whether the road is abnormal based on the first section data. Selecting the section data with the largest yaw rate as the first section data avoids analyzing the cause of yaw for multiple section data, reduces the amount of data to be processed for yaw cause analysis, and improves the speed of determining abnormal section data while reducing computing power occupation and energy consumption.

[0080] In one embodiment of the present invention, Figure 3 As shown, the road section data screening method includes:

[0081] Step S302, obtaining the trajectory data of historical orders;

[0082] Step S304, using a sequence mining algorithm to mine the trajectory data to determine multiple road section data;

[0083] Step S306, determining the yaw rate of a historical order under any road section data among the plurality of road section data;

[0084] Step S308, sorting the plurality of road section data according to the yaw rate, and selecting the road section data corresponding to the maximum yaw rate as the first road section data;

[0085] Step S310, classifying historical orders according to the first road section data;

[0086] Step S312, when the yaw rate of historical orders whose trajectory data does not include the first road section data is greater than the first yaw value, or when the number of historical orders whose trajectory data does not include the first road section data is greater than the first order value: select any one road section data from multiple road section data that do not include the first road section data as the second road section data.

[0087] In this embodiment, after determining the first section data, the first section data is used to classify historical orders to obtain historical orders containing the first section data and historical orders not containing the first section data. In order to eliminate the screening of later section data by the first section data, historical orders not containing the first section data are selected to perform a new round of screening of section data. Prior to this, it is determined whether the historical orders not containing the first section data meet the conditions for screening the section data again. Specifically, when the yaw rate of the historical orders whose trajectory data do not contain the first section data is less than the first yaw value, that is, the probability of yaw in the historical orders not containing the first section data is low, in order to reduce the occupation of computing power and energy consumption, No new round of screening of road section data is performed. Similarly, when the number of historical orders in which the trajectory data does not contain the first road section data is less than the first order value, the number of historical orders in which the trajectory data does not contain the first road section data is too small. In order to reduce the occupation of computing power and energy consumption, no new round of screening of road section data is performed. Only when the yaw rate of historical orders in which the trajectory data does not contain the first road section data is greater than the first yaw value, or when the number of historical orders in which the trajectory data does not contain the first road section data is greater than the first order value, one road section data is selected from the multiple road section data that do not contain the first road section data as the second road section data, and then the road section data reflecting the cause of the yaw is screened for analysis. The above steps of the technical solution can eliminate the influence of the first road section data on the screening of the second road section data, and the second road section data obtained can truly reflect the cause of the yaw, which is convenient for finding abnormal sections.

[0088] In one embodiment of the present invention, Figure 4 As shown, the road section data screening method includes:

[0089] Step S402, obtaining the trajectory data of historical orders;

[0090] Step S404, using a sequence mining algorithm to mine the trajectory data to determine multiple road section data;

[0091] Step S406, determining the yaw rate of a historical order under any road section data among the plurality of road section data;

[0092] Step S408, sorting the plurality of road section data according to the yaw rate, and selecting the road section data corresponding to the maximum yaw rate as the first road section data;

[0093] Step S410, classifying historical orders according to the first road section data;

[0094] Step S412, when the yaw rate of the historical order in which the trajectory data does not include the first road segment data is greater than the first yaw value, or when the number of historical orders in which the trajectory data does not include the first road segment data is greater than the first order value: select one road segment data from the multiple road segment data that do not include the first road segment data as the second road segment data;

[0095] Step S414, taking the historical orders that do not contain the first section data as the updated historical orders; classifying the updated historical orders according to the second section data until the yaw rate of the updated historical orders is less than the first yaw value or the number of updated historical orders is less than the first order value.

[0096] In this embodiment, by taking the historical order that does not contain the first road section data as the updated historical order, the previous judgment standard can be used in the new round of judgment process. Since the judgment conditions for judging whether a new round of road section data screening is required are the same, the road section data screening process can be performed multiple times until the conditions for further screening are not met. The above process is classified using the idea of ​​decision tree. Through execution, road section data that do not affect each other can be screened out so as to analyze the cause of deviation, avoiding the problem of needing to repeatedly analyze the same historical order multiple times at the current stage, which requires a large amount of data processing, and at the same time avoiding the road section data screened out being associated with each other, affecting the analysis of the cause of deviation.

[0097] In one embodiment of the present invention, Figure 5 As shown, the step of selecting one road section data from a plurality of road section data not including the first road section data as the second road section data specifically includes:

[0098] Step S502, sorting a plurality of road section data excluding the first road section data based on the yaw rate to obtain a sorting result, and selecting a preset number of road section data according to the sorting result;

[0099] Step S504: Select any one of the preset number of road section data as the second road section data.

[0100] In this embodiment, in order to improve the screening speed of the road section data, the number of multiple road section data is limited in the process of performing a new round of road section data screening. Further, the multiple road section data that do not include the first road section data are sorted based on the yaw rate, and a preset number of road section data are selected as candidate road section data according to the sorting result, and the second road section data are screened out based on the candidate road section data. Further, the preset number is the result of multiple experiments, and the candidate road section data obtained by screening is the road section data obtained after selecting the preset number from high to low in the sorting result.

[0101] In one embodiment of the present invention, Figure 6 As shown, the road section data screening method includes:

[0102] Step S602, obtaining the trajectory data of historical orders;

[0103] Step S604, using a sequence mining algorithm to mine the trajectory data to determine multiple road section data;

[0104] Step S606, determining the yaw rate of a historical order under any road section data among the plurality of road section data;

[0105] Step S608, sorting the plurality of road section data according to the yaw rate, and selecting the road section data corresponding to the maximum yaw rate as the first road section data;

[0106] Step S610, classifying historical orders according to the first road section data;

[0107] Step S612, when the yaw rate of the historical order in which the trajectory data does not include the first road segment data is greater than the first yaw value, or when the number of historical orders in which the trajectory data does not include the first road segment data is greater than the first order value: select one road segment data from the multiple road segment data that do not include the first road segment data as the second road segment data;

[0108] Step S614, when it is confirmed that an unordered set of road segment data obtained by filtering according to the yaw rate exists, the classification of historical orders is stopped.

[0109] In this embodiment, in order to further reduce the time spent on the screening process of the road segment data, it is determined whether an unordered set composed of the road segment data screened according to the yaw rate exists. If the determined unordered set exists, the classification of the historical orders is stopped. By determining whether the unordered set exists, if the unordered sets corresponding to the determined multiple road segment data already exist, the classification of the historical orders is stopped in advance, thereby avoiding the screening of the road segment data, reducing the amount of data to be processed and the energy consumption.

[0110] In one embodiment of the present invention, Figure 7 As shown, the steps of confirming the existence of the unordered set of selected road segment data specifically include:

[0111] Step S702, searching for a unique code corresponding to the road section data obtained by screening according to the yaw rate;

[0112] Step S704: if the unordered set consisting of unique codes exists, it is determined that the unordered set consisting of the road segment data obtained by screening according to the yaw rate exists.

[0113] In this embodiment, by searching for the unique code of the road section data, using the code to construct an unordered set, and determining whether the unordered set formed by the unique code exists. Specifically, the method for determining the unordered set formed by the unique code is simple, and the computing power required is low compared to the unordered set formed by directly using the selected road section data, and the amount of data processed is small, and thus, less time is required to obtain the same determination result, thereby improving the determination speed.

[0114] In one embodiment of the present invention, Figure 8 As shown, the road section data screening method includes:

[0115] Step S802, obtaining the trajectory data of historical orders;

[0116] Step S804, using a sequence mining algorithm to mine the trajectory data to determine multiple road section data;

[0117] Step S806, determining the yaw rate of a historical order under any road section data among the plurality of road section data;

[0118] Step S808, sorting the plurality of road section data according to the yaw rate, and selecting the road section data corresponding to the maximum yaw rate as the first road section data;

[0119] Step S810, classifying historical orders according to the first road section data;

[0120] Step S812, when the ratio of the number of historical orders whose trajectory data does not contain the road section data obtained by filtering the yaw rate to the number of historical orders is less than a third preset threshold, stop classifying the historical orders; or when the number of historical orders whose trajectory data does not contain the road section data obtained by filtering the yaw rate is less than a fourth preset threshold, stop classifying the historical orders.

[0121] In this embodiment, in order to further improve the screening speed of road section data, it is necessary to propose new pruning conditions for classifying historical orders. For example, when the ratio of the number of historical orders whose trajectory data does not contain road section data obtained by filtering according to the yaw rate to the number of historical orders is less than a third preset threshold, the classification of historical orders is stopped; or when the number of historical orders whose trajectory data does not contain road section data obtained by filtering according to the yaw rate is less than a fourth preset threshold, the classification of historical orders is stopped, thereby reducing the demand for computing power and energy consumption in the process of classifying historical orders.

[0122] In one embodiment of the present invention, the trajectory data includes the actual walking trajectory and the navigation trajectory, such as Fig. 9 As shown, the sequence mining algorithm is used to mine the trajectory data to determine the steps of multiple road segment data, specifically including:

[0123] Step S902, using a sequence mining algorithm to mine the actual walking trajectory and the navigation trajectory to determine a first group of road segment data sets and a second group of road segment data sets;

[0124] Step S904: determining a plurality of road section data according to the intersection of the first group of road section data set and the second group of road section data set.

[0125] In this embodiment, the trajectory data is historical travel data of online car-hailing vehicles, rental vehicles, private vehicles, etc., wherein the trajectory data includes two situations, the trajectory data is a navigation trajectory or the trajectory data is an actual travel trajectory. Therefore, when determining multiple road section data, firstly determine the first group of road section data sets according to whether the trajectory data is an actual travel trajectory, determine the second group of road section data sets according to whether the trajectory data is a navigation trajectory, and determine multiple road section data by the intersection of the first group of road section data sets and the second group of road section data sets; or first determine the second group of road section data sets according to whether the trajectory data is an actual travel trajectory, determine the first group of road section data sets according to whether the trajectory data is a navigation trajectory, and determine multiple road section data by the intersection of the first group of road section data sets and the second group of road section data sets, and filter the road section data according to the historical orders with deviation, and then analyze the reasons for the deviation.

[0126] In one embodiment of the present invention, the support degree of the first group of road section data and the second group of road section data is greater than a preset threshold; or the support degree of multiple road section data is greater than a preset threshold.

[0127] In this embodiment, in order to further improve the screening speed of road section data, while determining multiple road section data, the support of the sequence mining algorithm is set to filter out the road section data with a support less than a preset threshold, and only the road section data with a support greater than the preset threshold is screened, thereby reducing the number of multiple road section data, reducing the amount of data processing, and reducing energy consumption.

[0128] In one embodiment of the present invention, the sequence mining algorithm is a binary PrefixSpan algorithm.

[0129] In one embodiment of the present invention, the road segment data screening method includes: generating a candidate OD set and selecting a key OD.

[0130] The definitions of terms in the embodiments of the present invention include:

[0131] Road segment r: A directed edge in a road network. The two endpoints of road segment r are rs and re. Vehicles in the road network go from rs to re. The road network structure is expressed by the road segments and the topological relationships between them.

[0132] Trajectory traj: A trajectory is represented by a sequence of continuous road segments, for example: traj = r 1 →r2 →r 3 →...→r n , r (k+1) .s=r k .e,r (k+1) .s=r k .e, k∈[1,n-1]. The trajectories in this embodiment are divided into navigation trajectories and actual trajectories. Trajectory set TRAJ = {traj 1 ,traj 2 ,...,traj m}.

[0133] OD: Any road segment subsequence of length 2 that appears in the trajectory set TRAJ under study <r O ,r D >. OD support: sup <r O ,r D > is a subsequence in the trajectory set <r O ,r D >Number of occurrences.

[0134] Hot OD: OD with support greater than a certain threshold, currently set to 10. Equivalent to a frequent subsequence with a length of 2 and a support threshold of 10.

[0135] Deviation: refers to the behavior in which the actual track of the order is inconsistent with the navigation track on the target section.

[0136] Deviation rate: In the order trajectory set, the number of deviation orders / total number of orders.

[0137] First, the candidate OD set generation specifically includes:

[0138] The binary PrefixSpan algorithm is used to perform frequent subsequence mining on the trajectory data of historical orders, and only frequent sequences of length 2 are mined. The specific method is as follows: start mining sequence patterns from a prefix of length 1, search the corresponding projection database (suffix database) to obtain the frequent sequence corresponding to the prefix of length 1, and then recursively mine the frequent sequence of length 2.

[0139] Frequent sequence: a subsequence whose support is greater than a certain threshold.

[0140] The prefix is ​​a subsequence of the previous part of the sequence data. The mathematical description is as follows:

[0141] For the sequence A = {a 1 ,a 2 ,...a n} and sequence B = {b 1 ,b 2 ,...bm}, n≤m, satisfying a 1 =b 1 , a 2 =b 2 , ...a n =b n , then A is said to be a prefix of B.

[0142] In the PrefixSpan algorithm, the combination of all suffixes corresponding to the same prefix is ​​called the projection database corresponding to the prefix.

[0143] Specifically, as shown in Table 1 and Table 2, a, b, c, d, and e are elements in the sequence.

[0144] Table 1

[0145] ID (unique number of the sequence) Sequence 0 <abcd> < / abcd> 1 <bacd> < / bacd> 2 <eacba> < / eacba> 3 <bdaeb> < / bdaeb> 4 <dba> < / dba>

[0146] Table 2

[0147] A prefix of length 1 a b c d e Support 5 5 3 4 2

[0148] Assuming the support is 3, then d and e do not meet the requirements. In the following suffix database, they are deleted and no longer considered. For example, the sequence 3 <bdaeb>Change to <bdab>The obtained projection database is shown in Table 3.

[0149] Table 3

[0150] Prefix suffix a <bcd> ， <cd> ， <cba> ， < / cba> < / cd> < / bcd> b <cd> ， <acd> ， ， <ab> ，< / ab> < / acd> < / cd> c <d> ， <d> ， <ba> < / ba> < / d> < / d> d <aeb> ， <ba> < / ba> < / aeb>

[0151] Secondly, key OD selection includes:

[0152] By utilizing the principle of cutting all sample sets in the decision tree, multiple decision trees are constructed to search for key OD combinations that meet the conditions.

[0153] Construct a decision tree structure. The structure of the tree is as follows Fig.10 As shown, non-leaf nodes are represented by key OD (such as<O1,D1> ,<O2,D2> ,<O3,D3> ), and the leaf nodes are composed of order sets ({orders}). Specifically, the left side of each non-leaf node is a leaf node, which represents the set of all orders that have passed through the OD, and the right side is another non-leaf node or the set of orders that have not passed through the OD.

[0154] The tree structure construction process includes:

[0155] 1) Select a key OD from the set of ODs passed by the current set of all order trajectories, put all the orders that have passed through its OD into a set and put it into the left node, and put the set of orders that have not passed through the OD into the right node.

[0156] Order actual trajectory set: Orders are filtered through certain filtering conditions (currently all orders passing through a certain section of road are used in the experiment) to obtain an order trajectory set. All trajectories in the order set are the trajectory data we are concerned about, including actual trajectory data and navigation trajectory data. Orders are historical orders in this application.

[0157] 2) Check whether the right node meets the following conditions: the yaw rate is less than 0.1 or the number of order sets is less than 10. If so, the combination tree is constructed. If not, repeat the process 1) for the right node.

[0158] Pruning strategy:

[0159] 1) If the unordered set of all nodes on all current paths has appeared in the previous construction process, the tree construction will no longer be performed; specifically, for each OD sequence that has been passed, the serial number of the OD in the sequence is encoded into a unique integer through binary encoding and stored. For each OD sequence, it is encoded and checked whether it is in the encoded set. If so, the tree growth is stopped.

[0160] 2) If the order set passing through the current node meets the conditions: the yaw order is less than or equal to 5 or the yaw rate is less than or equal to 0.5, then a judgment is made. If the yaw order in the current order set is less than 15, the construction of a tree is completed, otherwise the current combination is discarded and the tree is no longer constructed.

[0161] OD Selection Process:

[0162] For the set of all ODs that the current order passes through, for each OD, the yaw rate and the number of yaw orders of all orders in the set that pass through the OD are calculated, and they are sorted according to the number of yaw orders and yaw rate, and the first n (multiple groups of experiments were conducted on the number of n) ODs are selected as the candidate set of nodes.

[0163] In an embodiment of the second aspect of the present invention, Fig.11 As shown, the road segment data screening system 1000 includes: a memory 1002, configured to store executable instructions; a processor 1004, configured to execute the stored instructions to: obtain trajectory data of historical orders; use a sequence mining algorithm to mine the trajectory data to determine multiple road segment data; determine the yaw rate of the historical order under any road segment data among the multiple road segment data; and screen the road segment data according to the yaw rate.

[0164] In this embodiment, the processor 1004 is configured to execute the executable instructions stored in the memory 1002 to: obtain the trajectory data of the historical orders, wherein the historical orders are historical travel data of online car-hailing, rental vehicles, private vehicles, etc., mine the trajectory data by using a sequence mining algorithm, determine multiple road segment data, select multiple road segments with a certain support from the trajectory data, wherein each of the multiple road segments has a key OD (Origin Destination, departure place and travel destination), but it should be noted that OD is only the departure place and travel destination of each section in the trajectory data of the historical order, rather than the departure place and destination of the entire historical order route, and the yaw rate of the historical order under each section data in the multiple section data is determined, and the section data is obtained by screening from the multiple section data according to the yaw rate, wherein the yaw rate is the ratio of the number of yaw orders in the historical orders passing through the section data to the historical orders, and the multiple section data are screened using the yaw rate to screen out the key section data that can reflect the yaw problem, thereby avoiding the inaccuracy of selecting key sections for yaw problem research only through road traffic, and the section data is obtained through screening for research, thereby improving the accuracy of the research results for the sections, wherein, by using the technical solution of the present application, compared with the method of determining key section data only through traffic in the related technical solution, there is no need to process the section data screened according to traffic one by one, thereby reducing the amount of data to be processed, thereby reducing the demand for computing power, and can select processing hardware with lower computing power to achieve hardware support, and at the same time, due to the reduction in processing volume, it can effectively reduce energy consumption.

[0165] In one embodiment of the present invention, Fig.11 As shown, the processor 1004 is configured to execute the executable instructions stored in the memory 1002 to: sort the multiple road segment data according to the yaw rate, and select the road segment data corresponding to the maximum yaw rate as the first road segment data.

[0166] In this embodiment, the processor 1004 is configured to execute the executable instructions stored in the memory 1002 to: screen multiple road sections by yaw rate, so that the yaw rate is used to sort the multiple road sections during the screening process. Since the first road section data is the road section data corresponding to the largest yaw rate, it can better reflect the yaw cause than other road section data in the process of studying the yaw problem, so it can be determined whether the road is abnormal based on the first road section data. The road section data with the largest yaw rate is selected as the first road section data, which avoids analyzing the yaw cause for multiple road section data, reduces the amount of data that needs to be processed for the yaw cause analysis, and improves the speed of determining the abnormal road section data while reducing the occupation of computing power and energy consumption.

[0167] In one embodiment of the present invention, Fig.11 As shown, the processor 1004 is also configured to execute executable instructions stored in the memory 1002 to: classify historical orders according to the first road section data; when the yaw rate of historical orders whose trajectory data does not contain the first road section data is greater than the first yaw value, or when the number of historical orders whose trajectory data does not contain the first road section data is greater than the first order value: select one road section data from multiple road section data that do not contain the first road section data as the second road section data.

[0168] In the technical solution, the processor 1004 is further configured to execute the executable instructions stored in the memory 1002 to: after determining the first road section data, use the first road section data to classify the historical orders to obtain the historical orders containing the first road section data and the historical orders not containing the first road section data; in order to eliminate the screening of the later road section data by the first road section data, select the historical orders not containing the first road section data to perform a new round of screening of the road section data; before this, determine whether the historical orders not containing the first road section data meet the conditions for screening the road section data again; specifically, when the yaw rate of the historical orders whose trajectory data do not contain the first road section data is less than the first yaw value, that is, the probability of yaw in the historical orders not containing the first road section data is low, In order to reduce the occupation of computing power and energy consumption, a new round of screening of road section data is no longer performed. Similarly, when the number of historical orders in which the trajectory data does not contain the first road section data is less than the first order value, the number of historical orders in which the trajectory data does not contain the first road section data is too small. In order to reduce the occupation of computing power and energy consumption, a new round of screening of road section data is no longer performed. Only when the yaw rate of the historical orders in which the trajectory data does not contain the first road section data is greater than the first yaw value, or when the number of historical orders in which the trajectory data does not contain the first road section data is greater than the first order value, one road section data is selected from the multiple road section data that do not contain the first road section data as the second road section data, and then the road section data reflecting the cause of the yaw is screened for analysis. The above steps of the technical solution can eliminate the influence of the first road section data on the screening of the second road section data, and then the obtained second road section data can truly reflect the cause of the yaw, which is convenient for finding abnormal sections.

[0169] In one embodiment of the present invention, Fig.11 As shown, the processor 1004 is also configured to execute executable instructions stored in the memory 1002 to: treat historical orders that do not contain the first road section data as updated historical orders; and classify the updated historical orders according to the second road section data until the yaw rate of the updated historical orders is less than the first yaw value or the number of updated historical orders is less than the first order value.

[0170] In this embodiment, the processor 1004 is further configured to execute the executable instructions stored in the memory 1002 to: by taking the historical order that does not contain the first road section data as the updated historical order, and then in the new round of judgment process, the previous judgment criteria can also be used. Since the judgment conditions for judging whether a new round of road section data screening is required are the same, the road section data screening process can be performed multiple times until the conditions for further screening are not met. The above process is classified using the idea of ​​decision tree, and through execution, road section data that do not affect each other can be screened out to analyze the cause of deviation, avoiding the problem of needing to repeatedly analyze the same historical order multiple times at this stage, and then requiring a large amount of data processing, and at the same time avoiding the road section data screened out being associated with each other, affecting the analysis of the cause of deviation.

[0171] In one embodiment of the present invention, Fig.11 As shown, the processor 1004 is configured to execute executable instructions stored in the memory 1002 to: sort multiple road section data that do not include the first road section data based on the yaw rate to obtain a sorting result, and select a preset number of road section data according to the sorting result; and select any one road section data from the preset number of road section data as the second road section data.

[0172] In this embodiment, in order to improve the screening speed of the road section data, the number of multiple road section data is limited during a new round of road section data screening, and the processor 1004 is configured to execute executable instructions stored in the memory 1002 to: sort the multiple road section data that do not include the first road section data based on the yaw rate, select a preset number of road section data as candidate road section data according to the sorting result, and screen out the second road section data based on the candidate road section data. Further, the preset number is the result of multiple experiments, and the candidate road section data obtained by screening is the road section data obtained after selecting the preset number from high to low in the sorting result.

[0173] In one embodiment of the present invention, Fig.11 As shown, the processor 1004 is further configured to execute executable instructions stored in the memory 1002 to: when it is confirmed that an unordered set of road segment data obtained by filtering according to the yaw rate exists, stop classifying the historical orders.

[0174] In this embodiment, in order to further reduce the time spent in the screening process of the road segment data, the processor 1004 is further configured to execute the executable instructions stored in the memory 1002 to: determine whether the unordered set composed of the road segment data screened according to the yaw rate exists, and if the determined unordered set exists, stop the classification of the historical orders. By determining whether the unordered set exists, if the unordered sets corresponding to the determined plurality of road segment data already exist, the classification of the historical orders is stopped in advance, thereby avoiding the screening of the road segment data, and reducing the amount of data to be processed and the energy consumption.

[0175] In one embodiment of the present invention, Fig.11 As shown, the processor 1004 is configured to execute the executable instructions stored in the memory 1002 to: search for the unique code corresponding to the road segment data obtained by filtering according to the yaw rate; if an unordered set consisting of the unique code exists, it is determined that the unordered set consisting of the road segment data obtained by filtering according to the yaw rate exists.

[0176] In this embodiment, the processor 1004 is configured to execute the executable instructions stored in the memory 1002 to: find the unique code of the road section data, use the code to build an unordered set, and determine whether the unordered set formed by the unique code exists. Specifically, the method for determining the unordered set formed by the unique code is simple, and the computing power required is low compared to the unordered set formed by directly using the selected road section data, and the amount of data processed is small, so that less time is required to obtain the same determination result, thereby improving the determination speed.

[0177] In one embodiment of the present invention, Fig.11 As shown, the processor 1004 is further configured to execute executable instructions stored in the memory 1002 to: stop classifying the historical orders when the ratio of the number of historical orders whose trajectory data does not contain the road segment data obtained by filtering the yaw rate to the number of historical orders is less than a third preset threshold, or stop classifying the historical orders when the number of historical orders whose trajectory data does not contain the road segment data obtained by filtering the yaw rate is less than a fourth preset threshold.

[0178] In this embodiment, in order to further improve the screening speed of the road segment data, it is necessary to propose new pruning conditions for classifying the historical orders. The processor 1004 is also configured to execute the executable instructions stored in the memory 1002 to: when the ratio of the number of historical orders whose trajectory data does not contain the road segment data obtained by filtering according to the yaw rate to the number of historical orders is less than a third preset threshold, stop classifying the historical orders; when the number of historical orders whose trajectory data does not contain the road segment data obtained by filtering according to the yaw rate is less than a fourth preset threshold, stop classifying the historical orders, thereby reducing the demand for computing power and energy consumption in the process of classifying the historical orders.

[0179] In one embodiment of the present invention, Fig.11 As shown, the trajectory data includes an actual walking trajectory and a navigation trajectory, and the processor 1004 is configured to execute executable instructions stored in the memory 1002 to: use a sequence mining algorithm to mine the actual walking trajectory and the navigation trajectory to determine a first set of road segment data sets and a second set of road segment data sets; determine multiple road segment data based on the intersection of the first set of road segment data sets and the second set of road segment data sets.

[0180] In this embodiment, the trajectory data is historical travel data of online car-hailing, rental vehicles, private vehicles, etc., wherein the trajectory data includes two situations, the trajectory data is a navigation trajectory or the trajectory data is an actual travel trajectory. The processor 1004 is configured to execute the executable instructions stored in the memory 1002 to: when determining multiple road segment data, first determine a first set of road segment data sets according to whether the trajectory data is an actual travel trajectory, determine a second set of road segment data sets according to whether the trajectory data is a navigation trajectory, and determine multiple road segment data by the intersection of the first set of road segment data sets and the second set of road segment data sets; or first determine a second set of road segment data sets according to whether the trajectory data is an actual travel trajectory, determine a first set of road segment data sets according to whether the trajectory data is a navigation trajectory, and determine multiple road segment data by the intersection of the first set of road segment data sets and the second set of road segment data sets, and filter the road segment data according to the historical orders with deviation, and then analyze the reasons for deviation.

[0181] In one embodiment of the present invention, Fig.11 As shown, the support degree of the first group of road section data and the second group of road section data is greater than a preset threshold; or the support degree of multiple road section data is greater than the preset threshold.

[0182] In this embodiment, in order to further improve the screening speed of road section data, while determining multiple road section data, the support of the sequence mining algorithm is set to filter out the road section data with a support less than a preset threshold, and only the road section data with a support greater than the preset threshold is screened, thereby reducing the number of multiple road section data, reducing the amount of data processing, and reducing energy consumption.

[0183] In one embodiment of the present invention, the sequence mining algorithm is a binary PrefixSpan algorithm.

[0184] In an embodiment of the third aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored, and the computer program is executed by a processor to implement any of the above-mentioned road section data screening methods.

[0185] The present invention provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above-mentioned road section data screening methods are implemented. Therefore, it has all the technical effects of any of the above-mentioned road section data screening methods, which will not be repeated here.

[0186] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.< / bdab> < / bdaeb>

Claims

1. A method for screening road section data, characterized in that: include: Get the trajectory data of historical orders; Using a sequence mining algorithm to mine the trajectory data to determine multiple road section data; Determine the yaw rate of the historical order under any road section data among the multiple road section data; sorting the plurality of road section data according to the yaw rate, and selecting the road section data corresponding to the maximum yaw rate as the first road section data; classifying the historical orders according to the first road section data; When the yaw rate of the historical orders in which the trajectory data does not include the first section data is greater than the first yaw value, or when the number of historical orders in which the trajectory data does not include the first section data is greater than the first order value: Select any one of the plurality of road section data not including the first road section data as the second road section data.

2. The method for selecting road section data according to claim 1, characterized in that: After the step of selecting one road section data from a plurality of road section data not including the first road section data as the second road section data, the method further includes: taking the historical orders not including the first road section data as updated historical orders; The updated historical orders are classified according to the second road section data until the yaw rate of the updated historical orders is less than the first yaw value or the number of the updated historical orders is less than the first order value.

3. The method for selecting road section data according to claim 1, characterized in that: The step of selecting one road section data from a plurality of road section data not including the first road section data as the second road section data specifically includes: sorting the plurality of road section data excluding the first road section data based on the yaw rate to obtain a sorting result, and selecting a preset number of road section data according to the sorting result; Select any one of the preset number of road section data as the second road section data.

4. The method for selecting road section data according to claim 1, characterized in that: Also includes: When it is confirmed that an unordered set of the road section data obtained by filtering according to the yaw rate exists, the classification of the historical orders is stopped.

5. The method for selecting road section data according to claim 4, characterized in that: The step of confirming the existence of the unordered set of selected road section data specifically includes: Finding a unique code corresponding to the road section data obtained by filtering the yaw rate; In the case that the unordered set composed of the unique codes exists, it is determined that the unordered set composed of the road section data obtained by screening according to the yaw rate exists.

6. The method for selecting road section data according to claim 1 or 4, characterized in that: Also includes: When the trajectory data does not contain a historical order of the road segment data obtained by filtering the yaw rate, and the ratio of the number of the historical orders to the historical orders is less than a third preset threshold, the classification of the historical orders is stopped; or When the number of historical orders in which the trajectory data does not include the road section data obtained by filtering according to the yaw rate is less than a fourth preset threshold, the classification of the historical orders is stopped.

7. The road section data screening method according to claim 1, characterized in that: The trajectory data includes a real-travel trajectory and a navigation trajectory. The step of mining the trajectory data using a sequence mining algorithm to determine a plurality of road segment data specifically includes: Using the sequence mining algorithm to mine the actual walking trajectory and the navigation trajectory, to determine a first group of road segment data sets and a second group of road segment data sets; The plurality of road segment data are determined according to an intersection of the first group of road segment data sets and the second group of road segment data sets.

8. The method for selecting road section data according to claim 7, characterized in that: The support degree of the first group of road section data and the second group of road section data is greater than a preset threshold; or the support degree of the plurality of road section data is greater than a preset threshold.

9. The method for selecting road section data according to claim 1, characterized in that: The sequence mining algorithm is a binary PrefixSpan algorithm.

10. A road section data screening system, characterized in that: include: a memory configured to store executable instructions; A processor configured to execute stored instructions to implement the steps of the road segment data screening method as described in any one of claims 1 to 9.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the road segment data screening method according to any one of claims 1 to 9 are implemented.

Images