CN110675074A - Travel target point identification method and device, and model development and evaluation method and device - Google Patents

Travel target point identification method and device, and model development and evaluation method and device Download PDF

Info

Publication number
CN110675074A
CN110675074A CN201910925488.1A CN201910925488A CN110675074A CN 110675074 A CN110675074 A CN 110675074A CN 201910925488 A CN201910925488 A CN 201910925488A CN 110675074 A CN110675074 A CN 110675074A
Authority
CN
China
Prior art keywords
target point
travel
target
identification model
point
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201910925488.1A
Other languages
Chinese (zh)
Inventor
查文斌
刘冬梅
张劲泉
赵琳
张晓亮
郭宇奇
侯德藻
汪林
王文静
王海鹏
乔国梁
王晶
丁丽媛
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Institute Of Highway Science Ministry Of Transport
Research Institute of Highway Ministry of Transport
Original Assignee
Institute Of Highway Science Ministry Of Transport
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Institute Of Highway Science Ministry Of Transport filed Critical Institute Of Highway Science Ministry Of Transport
Priority to CN201910994735.3A priority Critical patent/CN110728459A/en
Priority to CN201910925488.1A priority patent/CN110675074A/en
Publication of CN110675074A publication Critical patent/CN110675074A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/29—Geographical information databases
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G06N20/20—Ensemble learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/10—Services
    • G06Q50/26—Government or public services

Abstract

The invention provides a travel target point identification method and device, and a model development and evaluation method and device, wherein the travel target point identification model evaluation method mainly comprises the following steps: identifying the travel data according to a preset travel target point identification model to generate an identification conversion point sequence; acquiring a real conversion point sequence of the trip data; calculating preset model evaluation indexes according to the identification conversion point sequence and the real conversion point sequence, wherein the preset model evaluation indexes comprise target behavior starting and stopping time errors, target behavior duration errors, target behavior center point offset distances and accuracy rates; and determining an evaluation result of the travel target point identification model according to a preset model evaluation index. By implementing the method, the evaluation result not only reflects the accuracy, but also can reflect the error conditions of expansion, dislocation and the like on the time sequence between the conversion point sequence identified by the trip target point identification model and the real conversion point sequence, so the evaluation result is more practical.

Description

Travel target point identification method and device, and model development and evaluation method and device
Technical Field
The invention relates to the field of data processing, in particular to a travel target point identification method and device and a model development and evaluation method and device.
Background
Urban diseases such as traffic jam caused by unbalanced traffic demand and supply can cause unnecessary economic loss, and meanwhile, new challenges of traffic management face a lot, and no matter whether the traffic infrastructure construction, the traffic organization management and the traffic operation management need to analyze traffic demand, analyze travel characteristics and master traffic demand characteristics. Traffic behavior is a derived demand, people are always the main body of traffic, and the preference (attribute characteristics) of people is a factor for determining travel characteristics such as a traffic travel mode. Therefore, the characteristic analysis of the travel in transit is based on each traffic subject, and the travel characteristics of residents in the area can be accurately grasped only on the basis of detailed and accurate characteristic data of the individual travel activity chain, including specific information such as departure time, arrival time, transfer time, travel mode and the like in each travel process of the individual. Generally, an individual trip is not only dependent on a single transportation mode, but is completed by combining multiple transportation modes, in an actual environment, a departure time and an arrival time can be directly obtained through individual trip data, but transfer time and a trip transportation mode cannot be directly obtained from the individual trip data, a transportation mode conversion point divides the trip into a combination of the single trip modes, and the transportation mode conversion point identification is a basis for the transportation trip mode identification, the individual trip information acquisition and the transportation mode sharing rate research. Therefore, the extraction problem of the individual trip information is mainly focused on the identification of the transfer behavior/transfer point and the transportation mode between the transfer points in one trip of the individual. In the existing transfer behavior/transfer point identification method, identification is performed through a classification model, but when a verification set constructed based on sampling data is used for verifying and evaluating the constructed classification model, due to the fact that the type and the quantity of the sampling data are limited, the obtained evaluation result cannot effectively reflect the fitting degree of the output result of the classification model and the real situation.
Disclosure of Invention
Therefore, the technical problem to be solved by the present invention is that the model evaluation method in the prior art cannot effectively reflect the degree of conformity between the output result of the classification model and the actual situation, so as to provide a target point identification method and apparatus, and a model development and evaluation method and apparatus.
The invention provides a travel target point identification model evaluation method in a first aspect, which comprises the following steps: identifying the travel data according to a preset travel target point identification model to generate an identification conversion point sequence; acquiring a real conversion point sequence of the trip data; calculating preset model evaluation indexes according to the identification conversion point sequence and the real conversion point sequence, wherein the preset model evaluation indexes comprise target behavior starting and stopping time errors, target behavior duration errors, target behavior center point offset distances and accuracy rates; and determining an evaluation result of the travel target point identification model according to a preset model evaluation index.
Optionally, the step of calculating a preset model evaluation index according to the identified transition point sequence and the real transition point sequence includes: acquiring the recognition starting time and the recognition ending time of the target behavior according to the recognition conversion point sequence; acquiring the real starting time and the real ending time of the target behavior according to the real conversion point sequence; and determining the error of the starting and stopping time of the target behavior according to the identification starting time, the identification stopping time, the real starting time and the real stopping time.
Optionally, the step of calculating a preset model evaluation index according to the identified transition point sequence and the real transition point sequence further includes: calculating the recognition duration and the actual duration of the target behavior according to the recognition starting time, the recognition ending time, the real starting time and the real ending time; and determining the target behavior duration error according to the identification duration and the actual duration.
Optionally, the step of calculating a preset model evaluation index according to the identified transition point sequence and the real transition point sequence includes: acquiring the latitude and longitude of the central point of the identification conversion point sequence according to the identification conversion point sequence; acquiring the latitude and longitude of the central point of the real conversion point sequence according to the real conversion point sequence; and determining the target behavior center point offset distance according to the earth radius, the latitude and the longitude of the center point of the identification conversion point sequence and the latitude and the longitude of the center point of the real conversion point sequence.
Optionally, the step of determining an evaluation result of the travel target point identification model according to a preset model evaluation index includes: and if the target behavior start-stop moment error, the target behavior duration error, the target behavior center point offset distance and the accuracy are all larger than respective threshold values, judging that the preset travel target point identification model is qualified.
The second aspect of the present invention provides a travel target point identification model development method, including: acquiring a plurality of pieces of travel data, wherein the travel data comprise a development set and a verification set; extracting the characteristic value of each sampling point in the development set and the characteristic value of each sampling point in the verification set; developing an initial travel target point identification model according to a preset integrated learning method and the characteristic values of all sampling points in the training set; acquiring a real conversion point sequence according to the verification set; inputting the characteristic values of all sampling points in the verification set into an initial trip target point identification model to obtain an identification conversion point sequence; calculating preset model evaluation indexes according to the identification conversion point sequence and the real conversion point sequence, wherein the preset model evaluation indexes comprise target behavior starting and stopping time errors, target behavior duration errors, target behavior center point offset distances and accuracy rates; determining an evaluation result of the initial travel target point identification model according to a preset model evaluation index; if the evaluation result is qualified, determining the initial travel target point identification model as a travel target point identification model; and if the evaluation result is unqualified, returning to the step of extracting the characteristic value of each sampling point in the training set and the characteristic value of each sampling point in the verification set, or developing an initial trip target point identification model according to a preset integrated learning method and the characteristic values of each sampling point in the training set.
A third aspect of the present invention provides a travel target point identification method, including: acquiring outgoing data to be predicted; extracting characteristic values of sampling points in the row data to be predicted; and inputting the characteristic values of the sampling points in the outgoing data to be predicted into an outgoing target point identification model to obtain a target point sequence, wherein the outgoing target point identification model is obtained according to the outgoing target point identification model development method provided by the second aspect of the invention.
A fourth aspect of the present invention provides a travel target point identification model evaluation device, including: the system comprises a travel target point identification module, a travel conversion point sequence acquisition module and a travel conversion point sequence generation module, wherein the travel target point identification module is used for identifying travel data according to a preset travel target point identification model and generating an identification conversion point sequence; the real conversion point sequence acquisition module is used for acquiring a real conversion point sequence of the trip data; the preset model evaluation index calculation module is used for calculating preset model evaluation indexes according to the identification conversion point sequence and the real conversion point sequence, and the preset model evaluation indexes comprise target behavior starting and stopping time errors, target behavior duration errors, target behavior center point offset distances and accuracy rates; and the travel target point identification model evaluation module is used for determining the evaluation result of the travel target point identification model according to the preset model evaluation index.
A fifth aspect of the present invention provides a travel target point identification model development apparatus, including: the trip data acquisition module is used for acquiring a plurality of trip data, and the trip data comprises a development set and a verification set; the characteristic value extraction module is used for extracting and developing the characteristic values of the sampling points in the set and verifying the characteristic values of the sampling points in the set; the initial travel target point identification model establishing module is used for developing an initial travel target point identification model according to a preset integrated learning method and the characteristic values of all sampling points in the training set; the real conversion point sequence acquisition module is used for acquiring a real conversion point sequence according to the verification set; the identification conversion point sequence acquisition module is used for inputting the characteristic values of all the sampling points in the verification set into the initial trip target point identification model and acquiring an identification conversion point sequence; the preset model evaluation index calculation module is used for calculating preset model evaluation indexes according to the identification conversion point sequence and the real conversion point sequence, and the preset model evaluation indexes comprise target behavior starting and stopping time errors, target behavior duration errors, target behavior center point offset distances and accuracy rates; the initial travel target point identification model evaluation module is used for determining an evaluation result of the initial travel target point identification model according to a preset model evaluation index; and the trip target point identification model judging module is used for judging the initial trip target point identification model, if the evaluation result is qualified, the initial trip target point identification model is determined as the trip target point identification model, and if the evaluation result is unqualified, the execution characteristic value extracting module is triggered, or the initial trip target point identification model establishing module is triggered.
A sixth aspect of the present invention provides a travel target point recognition apparatus, including: the device comprises a to-be-predicted outgoing data acquisition module, a to-be-predicted outgoing data acquisition module and a to-be-predicted outgoing data acquisition module, wherein the to-be-predicted outgoing data acquisition module is used for acquiring outgoing data to be predicted; the system comprises a to-be-predicted row data characteristic value extraction module, a to-be-predicted row data characteristic value extraction module and a to-be-predicted row data characteristic value extraction module, wherein the to-be-predicted row data characteristic value extraction module is used for extracting the characteristic values of all sampling points in the to-be-predicted; and the target point sequence identification module is used for inputting the characteristic values of all sampling points in the outgoing data to be predicted into the outgoing target point identification model to obtain a target point sequence, and the outgoing target point identification model is obtained according to the outgoing target point identification model development method provided by the second aspect of the invention.
A seventh aspect of the present invention provides a computer apparatus comprising: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executable by the at least one processor to perform the travel target point recognition model evaluation method provided by the first aspect of the present invention, or the travel target point recognition model development method provided by the second aspect of the present invention, or the travel target point recognition method provided by the third aspect of the present invention.
An eighth aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions for causing a computer to execute the travel target point identification model evaluation method provided by the first aspect of the present invention, or the travel target point identification model development method provided by the second aspect of the present invention, or the travel target point identification method provided by the third aspect of the present invention.
The technical scheme of the invention has the following advantages:
1. according to the method for evaluating the travel target point identification model, when the travel target point identification model is evaluated, the target behavior start-stop time error, the target behavior duration error and the target behavior center point offset distance are added on the basis of the accuracy of the preset model evaluation index, and the travel target point identification model is evaluated through the method provided by the invention, so that the evaluation result not only reflects the accuracy, but also can reflect the expansion and dislocation and other error conditions of the time sequence between the conversion point sequence identified through the travel target point identification model and the real conversion point sequence, and therefore, the evaluation result is more practical.
2. The invention provides a travel target point identification model development method, after developing and generating an initial travel target point identification model through a training set and a preset integrated learning method, a verification set is input into the initial travel target point identification model to generate an identification conversion point sequence, a start-stop time error, a target behavior duration error, a target behavior center point offset distance and an accuracy rate are calculated according to the identification conversion point sequence and a real conversion point sequence, the initial travel target point identification model is evaluated through the start-stop time error, the target behavior duration error, the target behavior center point offset distance and the accuracy rate, if the evaluation result is qualified, the initial travel target point identification model is determined as a travel target point identification model, and if the evaluation result is unqualified, the characteristic values of each sampling point in the training set and the characteristic values of each sampling point in the verification set are returned and extracted, or, according to the preset integrated learning method and the characteristic values of all sampling points in the training set, the step of developing the initial trip target point identification model is developed again, and the trip target point identification model obtained by implementing the trip target point identification model development method provided by the invention effectively controls the errors such as expansion, dislocation and the like on the time sequence between the conversion point sequence identified by the trip target point identification model and the real conversion point sequence on the premise of ensuring the accuracy, so that the limitation of the existing trip target point identification model is overcome, and the accuracy is higher.
3. According to the travel target point identification method provided by the invention, when the target point sequence is identified, the adopted travel target point identification model is obtained through the travel target point identification model development method provided by the second aspect of the invention, so that the travel target point identification model has higher precision, and the target point sequence identified by the travel target point identification method provided by the invention is more accurate.
4. When the trip target point identification model is evaluated, the target behavior start-stop time error, the target behavior duration error and the target behavior center point offset distance are newly added on the basis of the accuracy of the used preset model evaluation index, and the evaluation result not only reflects the accuracy, but also can reflect the expansion and dislocation and other error conditions of the time sequence between the conversion point sequence identified by the trip target point identification model and the real conversion point sequence, so that the evaluation result is more practical.
5. The invention provides a travel target point identification model development device, after developing a preset neural network through a development set to generate an initial travel target point identification model, a verification set is also input into the initial travel target point identification model to generate an identification conversion point sequence, a start-stop time error, a target behavior duration error, a target behavior center point offset distance and an accuracy rate are calculated according to the identification conversion point sequence and a real conversion point sequence, the initial travel target point identification model is evaluated through the start-stop time error, the target behavior duration error, the target behavior center point offset distance and the accuracy rate, if the evaluation result is qualified, the initial travel target point identification model is determined as a travel target point identification model, if the evaluation result is unqualified, an execution characteristic value extraction module is triggered, or an initial travel target point identification model establishment module is triggered, the travel target point identification model developed by the travel target point identification model development device effectively controls the errors such as expansion, dislocation and the like on the time sequence between the conversion point sequence identified by the travel target point identification model and the real conversion point sequence on the premise of ensuring the accuracy, overcomes the limitation of the existing travel target point identification model, and has higher accuracy.
6. When the travel target point recognition device provided by the invention recognizes the target point sequence, the adopted travel target point recognition model is obtained by the development method of the travel target point recognition model provided by the second aspect of the invention, so that the travel target point recognition model has higher precision, and the target point sequence recognized by the travel target point recognition device provided by the invention is more accurate.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a diagram of a comparison of a sequence of transition points identified by the prior art with a sequence of true transition points;
fig. 2 to fig. 6 are flowcharts of specific examples of the travel target point identification model evaluation method according to the embodiment of the present invention;
fig. 7 is a flowchart illustrating a specific example of a travel target point identification model development method according to an embodiment of the present invention;
FIG. 8 is a graph illustrating a distribution analysis of model input features according to an embodiment of the present invention;
FIG. 9 is a graph of performance verification effects of the Random Forest model in an embodiment of the present invention;
FIG. 10 is a diagram illustrating the performance verification effect of the Adaboost model in an embodiment of the present invention;
FIG. 11 is a graph showing the performance verification effect of the Gradient Boosting Decision Tree model according to the embodiment of the present invention;
FIG. 12 is a diagram illustrating the performance verification effect of the XGboost model in an embodiment of the present invention;
fig. 13 is a flowchart of a specific example of a travel target point identification method according to an embodiment of the present invention;
fig. 14 is a block diagram illustrating a specific example of a trip target point identification model evaluation apparatus according to an embodiment of the present invention;
fig. 15 is a block diagram illustrating a specific example of a trip target point identification model development apparatus according to an embodiment of the present invention;
fig. 16 is a block diagram illustrating a specific example of a travel target point recognition apparatus according to an embodiment of the present invention;
fig. 17 is a block diagram showing a specific example of a computer device in the embodiment of the present invention.
Detailed Description
The technical solutions of the present invention will be described clearly and completely with reference to the accompanying drawings, and it should be understood that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In addition, the technical features involved in the different embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.
Example 1
In the existing method for identifying the transition points in the travel process, a method which is more accurate and rapid is to identify through a classification model, but when the transition points are identified through the classification model, errors usually occur between a sequence formed by the identified transition points and a sequence formed by real transition points, as shown in fig. 1, two sequences of the transition points identified through the classification model are listed, the identified sequence of the transition points is compared with the real sequence of the transition points, and as can be seen from fig. 1, the dislocation and the expansion existing on a time sequence between the identified sequence of the transition points and the real sequence of the transition points identified through the existing model are shown.
The embodiment of the invention provides a travel target point identification model evaluation method, as shown in fig. 2, comprising the following steps:
step S110: the travel data is identified according to a preset travel target point identification model to generate an identification conversion point sequence, in a specific embodiment, the preset travel target point identification model may be a classification model established by an integrated learning method such as adaptive Boosting-AdaBoost, Gradient Boosting Tree (Gradient Boosting Decision Tree), eXtreme Gradient Boosting-XGBoost, Random Forest (Random Forest) and the like, when identifying the travel data, each sampling point of the travel data is identified, whether each sampling point is a conversion point is respectively determined, so as to establish the identification conversion point sequence according to the identified continuous conversion points, wherein the conversion points refer to sampling points in the process of converting individual travel data from one transportation mode to another transportation mode, the research of the embodiment of the invention aims at identifying the conversion points, the target point described in the embodiments of the present invention is therefore referred to as a transition point.
Step S120: in a specific embodiment, the travel data is composed of a plurality of sampling points, the sampling points in the process of converting the individual travel data from one transportation mode to another transportation mode in the travel data are real conversion points, and the real conversion point sequence is a sequence established according to continuous real conversion points.
Step S130: and calculating preset model evaluation indexes according to the identification conversion point sequence and the real conversion point sequence, wherein the preset model evaluation indexes comprise target behavior start-stop moment errors, target behavior duration errors, target behavior center point offset distances and accuracy rates.
In one embodiment, the accuracy is calculated by the following equation (1):
wherein tp represents the real converting point, the identification result is the number of sampling points of the converting point, tn represents the real converting point, the identification result is the number of sampling points of the non-converting point, fp represents the real non-converting point, the identification result is the number of sampling points of the non-converting point, fn represents the real non-converting point, and the identification result is the number of sampling points of the converting point.
Step S140: and determining an evaluation result of the travel target point identification model according to a preset model evaluation index.
According to the method for evaluating the travel target point identification model provided by the embodiment of the invention, when the travel target point identification model is evaluated, the target behavior start-stop time error, the target behavior duration error and the target behavior center point offset distance are newly added on the basis of the accuracy of the used preset model evaluation index, and the travel target point identification model is evaluated by the method provided by the embodiment of the invention, so that the evaluation result not only reflects the accuracy, but also can reflect the error conditions of expansion, dislocation and the like on the time sequence between the travel target point identification model identification conversion point sequence and the real conversion point sequence, and therefore, the evaluation result is more practical.
In an alternative embodiment, as shown in fig. 3, step S130 specifically includes:
step S131: in a specific embodiment, the target behavior is a transfer behavior, and the recognition transition point sequence is constructed according to the continuously recognized transition points, so that the recognition start time of the target behavior is the time corresponding to the first transition point in the recognition transition point sequence, and similarly, the recognition end time of the target behavior is the time corresponding to the last transition point in the recognition transition point sequence.
Step S132: and acquiring the real starting time and the real ending time of the target behavior according to the real conversion point sequence, wherein in a specific embodiment, the real starting time of the target behavior is the time corresponding to the first conversion point in the real conversion point sequence, and the real ending time of the target behavior is the time corresponding to the last conversion point in the real conversion point sequence.
Step S133: and determining the error of the starting and stopping time of the target behavior according to the identification starting time, the identification stopping time, the real starting time and the real stopping time.
In a specific embodiment, the target behavior start-stop time error is calculated by the following equations (2) and (3):
Figure BDA0002218794800000081
where TimeDiff1 denotes the time error, Treal_startIndicating the real start time, Tidentify_startIndicates the recognition start time, Treal_endIndicating the true termination time, Tidentify_endIndicating the moment of termination of recognition.
Figure BDA0002218794800000082
WhereinPSTE2 denotes target behavior start-stop time error NUMTimeDiff1<2Represents a sum of time errors, NUM, of less than 2 minutesTimeDiff1In this embodiment, when calculating the target behavior start-stop time error, 2 minutes is used as an allowable error, and the percentage of the transfer time error in 2 minutes in one trip is used as an evaluation index.
In an alternative embodiment, as shown in fig. 4, step S130 further includes:
step S134: and calculating the recognition duration and the actual duration of the target behavior according to the recognition starting time, the recognition ending time, the real starting time and the real ending time.
Step S135: and determining the target behavior duration error according to the identification duration and the actual duration.
In one embodiment, the target behavior duration error is calculated by the following equations (4) and (5):
TimeDiff2＝||Treal_start-Treal_end|-|Tidentify_start-Tidentify_end||， (4)
where TimeDiff2 represents the duration error.
Figure BDA0002218794800000083
Where PTE2 represents target behavior duration error, NUMTimeDiff2<2Indicating a time error sum, NUM, of less than 2 minutesTimeDiff2In this embodiment, when calculating the target behavior duration error, 2 minutes is used as an allowable error, and the percentage of the transfer duration error in 2 minutes in the one-trip process is used as an evaluation index, and in practical application, different allowable errors may be set according to actual requirements.
In the embodiment of the invention, the target behavior start-stop time error and the target behavior duration error are added in the preset model evaluation index, and the expansion error condition of the time sequence between the travel target point identification model identification conversion point sequence and the real conversion point sequence can be reflected through the two evaluation indexes, so that the evaluation result is more practical.
In an alternative embodiment, as shown in fig. 5, step S130 includes:
step S136: and acquiring the latitude and longitude of the central point of the identification conversion point sequence according to the identification conversion point sequence.
Step S137: and acquiring the latitude and longitude of the central point of the real conversion point sequence according to the real conversion point sequence.
Step S138: and determining the target behavior center point offset distance according to the earth radius, the latitude and the longitude of the center point of the identification conversion point sequence and the latitude and the longitude of the center point of the real conversion point sequence.
In a specific embodiment, the target line center point offset distance is calculated by the following equations (6) and (7):
Figure BDA0002218794800000091
where DistDiff denotes offset distance, r denotes earth radius, latrealLatitude, long, representing the center point of the sequence of true transition pointsrealLongitude, lat, representing the center point of the real sequence of translation pointsidentifyIndicating the latitude, long, of the center point identifying the sequence of transition pointsidentifyIndicating the longitude of the center point identifying the sequence of transition points.
Figure BDA0002218794800000092
Where PCO 30' represents the target line center point offset distance, NUMDistDiff<30Denotes an offset distance sum, NUM, of less than 30 metersDistDiffThe total offset distance is shown, in this embodiment, 30 meters is used as an allowable error, and the percentage of the offset distance error of the center point within 30 meters in one trip is used as an evaluation index.
In the embodiment of the invention, the deviation distance representing the target behavior center point is added in the preset model evaluation index, and the error conditions such as expansion, dislocation and the like on the time sequence between the travel target point recognition model recognition conversion point sequence and the real conversion point sequence can be reflected by representing the target behavior center point deviation distance, so that the evaluation result is more practical.
In an alternative embodiment, as shown in fig. 6, step S140 specifically includes:
step S141: and if the target behavior start-stop moment error, the target behavior duration error, the target behavior center point offset distance and the accuracy are all larger than respective threshold values, judging that the preset travel target point identification model is qualified. And if any one of the target behavior starting and stopping time error, the target behavior duration error, the target behavior center point offset distance and the accuracy is smaller than the corresponding self-threshold value, judging that the preset travel target point identification model is unqualified.
Example 2
An embodiment of the present invention provides a travel target point identification model development method, as shown in fig. 7, including:
step S210: the method comprises the steps of obtaining a plurality of pieces of travel data, wherein the travel data comprise a training set and a verification set, and each piece of travel data is composed of a plurality of sampling points.
Step S220: and extracting the characteristic value of each sampling point in the training set and the characteristic value of each sampling point in the verification set.
In a specific embodiment, the feature value of each sampling point is calculated based on a sampling point sequence formed by N time length windows before and after the sampling point, and when the feature values of N sampling points before and after the trip data do not satisfy the window calculation condition, the feature value of each sampling point cannot be calculated. The features extracted for each sample point are shown in table 1 below:
TABLE 1
Figure BDA0002218794800000101
In one embodiment, the model input feature distribution analysis is shown in FIG. 8.
Step S230: and developing an initial travel target point identification model according to a preset integrated learning method and the characteristic values of all sampling points in the training set. In a specific embodiment, the preset ensemble learning method may be any one or more of Adaptive Boosting (AdaBoost), Gradient Boosting Tree (Gradient Boosting Decision Tree), eXtreme Gradient Boosting (XGBoost), Random Forest (Random Forest) and other ensemble learning methods.
Step S240: acquiring a real conversion point sequence according to the verification set;
step S250: inputting the characteristic values of all sampling points in the verification set into an initial trip target point identification model to obtain an identification conversion point sequence;
step S260: a preset model evaluation index is calculated according to the identified transition point sequence and the real transition point sequence, where the preset model evaluation index includes a target behavior start-stop time error, a target behavior duration error, a target behavior center point offset distance, and an accuracy rate, and the detailed description is described in the above embodiment 1 with respect to step S130.
Step S270: the evaluation result of the initial travel target point identification model is determined according to the preset model evaluation index, and the detailed description is given in the above embodiment 1 to the description of step S140.
Step S280: if the evaluation result is qualified, determining the initial travel target point identification model as a travel target point identification model; and if the evaluation result is unqualified, returning to the step of extracting the characteristic value of each sampling point in the training set and the characteristic value of each sampling point in the verification set, or developing an initial trip target point identification model according to a preset integrated learning method and the characteristic values of each sampling point in the training set.
The method for developing a trip target point identification model provided by the embodiment of the invention is characterized in that after an initial trip target point identification model is developed according to the preset integrated learning method and the characteristic values of all sampling points in a training set, a verification set is input into the initial trip target point identification model to generate an identification conversion point sequence, a start-stop time error, a target behavior duration error, a target behavior central point offset distance and an accuracy rate are calculated according to the identification conversion point sequence and a real conversion point sequence, the initial trip target point identification model is evaluated according to the start-stop time error, the target behavior duration error, the target behavior central point offset distance and the accuracy rate, if the evaluation result is qualified, the initial trip target point identification model is determined as the trip target point identification model, and if the evaluation result is unqualified, the extraction of the characteristic values of all sampling points in the training set and the characteristic values of all sampling points in the verification set are, or, the step of developing the initial trip target point identification model is carried out again according to the preset integrated learning method and the characteristic values of the sampling points in the training set, and the trip target point identification model developed by implementing the trip target point identification model development method provided by the embodiment of the invention effectively overcomes the errors of expansion, dislocation and the like on the time sequence between the conversion point sequence identified by the trip target point identification model and the real conversion point sequence on the premise of ensuring the accuracy, overcomes the limitation of the existing trip target point identification model, and has higher accuracy.
In a specific embodiment, if the step S220 is returned, the features are selected again, the feature values are extracted, and the initial travel target point identification model is developed according to the new feature values and a preset ensemble learning method. If the step S230 is returned, a new ensemble learning method needs to be selected again to develop the initial travel target point identification model based on the existing feature values.
In an optional embodiment, after step S210 and before step S220, the method for developing a travel target point identification model further includes:
step S290: and cleaning the trip data. In an embodiment, in step S280, when the evaluation result of the initial travel target point identification model is not qualified, the process may also return to step S290 to re-clean the data.
In one embodiment, the cleansing trip data includes the following four cases:
1. the trip data is acquired based on a GPS, and due to the fact that GPS signals are interfered and inevitable factors such as errors exist in equipment, the trip data generated by sampling are not completely the same as real trip data, and false recording points deviating from normal positions, namely abnormal points, exist in sampling points of the trip data. In the embodiment of the invention, the point with the instantaneous speed of more than 60km/h is regarded as an abnormal point, deleted and treated as a missing record;
2. taking the trip data with the duration less than 5min or the number of sampling points less than 60 as invalid trip data, and deleting the invalid trip data;
3. in the same trip mode segment, if the trip data missing time is longer than 2min, performing trip segment division processing, and enabling the former point to be the end point of the current trip and the latter point to be the starting point of the next trip; if the missing time is less than 2min, performing interpolation on the missing data by using Kalman filtering;
4. the identification of the abnormal points removes points with larger drift degree, but the remaining smaller drift still remains in the outgoing data, which is represented as a jump phenomenon of the outgoing data body within a certain radius range of the real position. The embodiment of the invention uses the extended Kalman filtering to correct the error of the track.
In a specific embodiment, the characteristic values of the sampling points in the training set can be developed through any one of the integrated learning methods such as adaptive enhancement, gradient lifting tree, extreme gradient lifting, random forest and the like, so as to obtain an initial travel target point identification model. However, in order to select an optimal travel target point identification model, in the embodiment of the present invention, feature values of sampling points in a training set are developed by adaptive enhancement, gradient lifting tree, extreme gradient lifting, and random forest ensemble learning methods, respectively, to obtain four initial travel target point identification models, and then an optimal model is selected from the four initial travel target point identification models as a travel target point identification model, where the specific development method is as follows:
the optimal hyper-parameter combination form of the AdaBoost model, the Random Forest model and the Gradient Boosting Decision Tree model is as follows: [ Window size (Window _ size), maximum depth (Max _ depth), leaf node minimum sample number (Min _ samples _ leaf), split node minimum sample number (Min _ samples _ split), number of trees in the forest (N _ estimators) ], for the XGBoost model, the optimal hyper-parameter combination is: window size (Window _ size), maximum depth (Max _ depth), subsample (subsample), subsample ratio (subsample _ byte) of the constructed tree, number of trees in the forest (N _ estimators) ]. Because the distribution of the feature data changes due to different sizes of the windows, the embodiment of the invention firstly determines the size of the window, and then optimizes the optimal hyper-parameters of the 4 models by using grid search on the basis of the size of the window, wherein the number of the windows and the hyper-parameter set are shown in the following table 2:
TABLE 2
Figure BDA0002218794800000131
The four models are developed through the steps S210 to S230, and the developed models are evaluated in combination with the steps S240 to S280, and the performance verification effect graphs and analysis of the models are shown in fig. 9 to 12.
As can be seen from fig. 9-12, the Random Forest model is optimized when the window size is 25, the Adaboost model and the Gradient Boosting Decision Tree (GBDT) model are optimized when the window size is 30, and the XGBoost model is optimized when the window size is 35, which indicates that the size of the window size can actually cause a change in data distribution, thereby affecting the performance of the model, and it is noted that even in the case of the model being optimized, the ratio of the start-stop time errors in the four models within 2min is very low, which causes this phenomenon because many respondents delay mark the transfer waiting state or end the transfer waiting state in advance, which results in identifying that the transfer point sequence is stretched in the time sequence compared with the real transfer point sequence, and the center transfer points are not far apart.
The method comprises the steps of selecting 22 sections of travel data which are independent by one person and take 12 days as test set data, and using the test set data to simulate errors of a travel target point identification model developed by the travel target point identification model development method provided by the embodiment of the invention in actual engineering and verify the effectiveness of the model. The preset model evaluation indexes of the different models are calculated through the above steps S240 to S270, the final results of the models are shown in the following table 3,
TABLE 3
Model name Optimal hyper-parametric combinations Accuracy(％) PSTE2(％) PTE2(％) PCO30(％)
Adaboost [30,20,4,5,500] 98.7 54.6 100 89.4
GBDT [30,20,8,5,300] 98.4 55.0 100 94.7
XGBoost [35,30,0.5,1,500] 98.3 52.6 100 93.8
Random Forest [25,20,4,5,400] 95.9 52.6 100 91.1
From the above four preset model evaluation indexes, although the Accuracy (Accuracy) of the GBDT model is slightly worse than that of Adaboost, the three model evaluation index values of the target behavior start-stop time error (PSTE2), the target behavior duration error (PTE2) and the target behavior center point offset distance (PCO30) are all the best of the four models, so that the GBDT model with the hyper-parameter combination of [30,20,8,5,300] is selected as the final travel target point identification model.
In the embodiment of the present invention, four models are developed according to the above steps S210 to S280, respectively, to obtain four initial travel target point identification models, and then an optimal travel target point identification model is selected from the four initial travel target point identification models, because the models have different performances, even if different models are developed according to the same steps, the obtained initial travel target point identification models have different performances.
Example 3
An embodiment of the present invention provides a travel target point identification method, as shown in fig. 13, including:
step S310: acquiring outgoing data to be predicted, wherein each piece of outgoing data consists of a plurality of sampling points;
step S320: the characteristic values of the sampling points in the row data to be predicted are extracted, and the detailed description is given in the above embodiment 2 for the description of step S220.
Step S330: the characteristic values of the sampling points in the outgoing data to be predicted are input into the trip target point identification model, a target point sequence is obtained, the trip target point identification model is obtained according to the trip target point identification model development method provided in the above embodiment 2, and the detailed description is given in the above embodiment 2 of the trip target point identification model development method.
According to the travel target point identification method provided by the embodiment of the invention, when the target point sequence is identified, the adopted travel target point identification model is obtained by the travel target point identification model development method provided by the embodiment 2 of the invention, so that the travel target point identification model has higher precision, and the target point sequence identified by the travel target point identification method provided by the invention is more accurate.
Example 4
An embodiment of the present invention provides a travel target point identification model evaluation device, as shown in fig. 14, including:
the identifying and converting point sequence obtaining module 110 is configured to identify the travel data according to a preset travel target point identifying model, and generate an identifying and converting point sequence, which is described in detail in the foregoing embodiment 1 for the step S110.
The real conversion point sequence obtaining module 120 is configured to obtain a real conversion point sequence of the trip data, and the detailed description is given in the above description of step S120 in embodiment 1.
The preset model evaluation index calculation module 130 is configured to calculate a preset model evaluation index according to the identified transition point sequence and the real transition point sequence, where the preset model evaluation index includes a target behavior start-stop time error, a target behavior duration error, a target behavior center point offset distance, and an accuracy rate, and the detailed description is described in the above embodiment 1 for step S130.
The travel target point identification model evaluation module 140 is configured to determine an evaluation result of the travel target point identification model according to a preset model evaluation index, which is described in detail in the above description of step S140 in embodiment 1.
When the trip target point identification model evaluation device provided by the embodiment of the invention evaluates the trip target point identification model, the target behavior start-stop time error, the target behavior duration error and the target behavior center point offset distance are newly added on the basis of the accuracy of the used preset model evaluation indexes, and the evaluation result not only reflects the accuracy, but also can reflect the expansion and dislocation and other error conditions of the time sequence between the conversion point sequence identified by the trip target point identification model and the real conversion point sequence, so that the evaluation result is more practical.
Example 5
An embodiment of the present invention provides a travel target point identification model development apparatus, as shown in fig. 15, including:
the travel data acquiring module 210 is configured to acquire a plurality of pieces of travel data, where the travel data includes a training set and a verification set, and the detailed description is described in the foregoing description of the step S210 in embodiment 2.
The feature value extracting module 220 is configured to extract feature values of the sampling points in the training set and feature values of the sampling points in the verification set, and the detailed description is described in the above embodiment 2 for the step S220.
The initial travel target point identification model establishing module 230 is configured to develop an initial travel target point identification model according to a preset ensemble learning method and a feature value of each sample point in a training set, and the detailed description is described in the above embodiment 2 for the step S230.
A real transformation point sequence obtaining module 240, configured to obtain a real transformation point sequence according to the verification set, which is described in detail in the above description of step S240 in embodiment 2.
The identifying and converting point sequence obtaining module 250 is configured to input the feature values of the sampling points in the verification set into the initial trip target point identifying model, and obtain an identifying and converting point sequence, which is described in detail in the above embodiment 2 for the step S250.
The preset model evaluation index calculation module 260 is configured to calculate a preset model evaluation index according to the identified transition point sequence and the real transition point sequence, where the preset model evaluation index includes a target behavior start-stop time error, a target behavior duration error, a target behavior center point offset distance, and an accuracy rate, and the detailed description is described in the above embodiment 2 for step S260.
The initial travel target point identification model evaluation module 270 is configured to determine an evaluation result of the initial travel target point identification model according to a preset model evaluation index, which is described in detail in the above description of the step S270 in embodiment 2.
The trip target point identification model determining module 280 is configured to determine the initial trip target point identification model, determine the initial trip target point identification model as the trip target point identification model if the evaluation result is qualified, and trigger the execution characteristic value extracting module if the evaluation result is unqualified, or trigger the initial trip target point identification model establishing module, which is described in detail in the description of step S280 in embodiment 2.
The trip target point identification model development device provided by the embodiment of the invention is characterized in that after an initial trip target point identification model is developed according to the preset integrated learning method and the characteristic values of sampling points in a training set, a verification set is input into the initial trip target point identification model to generate an identification conversion point sequence, a start-stop time error, a target behavior duration error, a target behavior central point offset distance and an accuracy rate are calculated according to the identification conversion point sequence and a real conversion point sequence, the initial trip target point identification model is evaluated according to the start-stop time error, the target behavior duration error, the target behavior central point offset distance and the accuracy rate, if the evaluation result is qualified, the initial trip target point identification model is determined as the trip target point identification model, and if the evaluation result is unqualified, a characteristic value extraction module is triggered, or, the initial trip target point identification model building module re-develops the trip target point identification model, and the trip target point identification model developed by the trip target point identification model development device provided by the embodiment of the invention effectively controls the errors such as expansion, dislocation and the like on the time sequence between the conversion point sequence identified by the trip target point identification model and the real conversion point sequence on the premise of ensuring the accuracy, thereby overcoming the limitation of the existing trip target point identification model and having higher accuracy.
Example 6
An embodiment of the present invention provides a travel target point identification device, as shown in fig. 16, including:
the to-be-predicted outgoing data acquiring module 310 is configured to acquire the to-be-predicted outgoing data, and the detailed description is provided in the above description of the step S310 in embodiment 3.
The module 320 for extracting the feature value of the row data to be predicted is used to extract the feature value of each sampling point in the row data to be predicted, and the detailed description is described in the above embodiment 3 for the step S320.
The target point sequence identification module 330 is configured to input the feature values of the sampling points in the outgoing line data to be predicted into the outgoing target point identification model, so as to obtain a target point sequence, where the outgoing target point identification model is obtained according to the method for developing the outgoing target point identification model provided in embodiment 2, and the detailed description is described in step S210 to step S280 in embodiment 2.
When the travel target point recognition device provided by the embodiment of the invention recognizes the target point sequence, the adopted travel target point recognition model is obtained by the travel target point recognition model development method provided by the embodiment 2 of the invention, so that the travel target point recognition model has higher precision, and the target point sequence recognized by the travel target point recognition device provided by the embodiment of the invention is more accurate.
Example 7
An embodiment of the present invention provides a computer device, as shown in fig. 17, the computer device mainly includes one or more processors 41 and a memory 42, and fig. 17 takes one processor 41 as an example.
The computer device may further include: an input device 43 and an output device 44.
The processor 41, the memory 42, the input device 43 and the output device 44 may be connected by a bus or other means, and the bus connection is exemplified in fig. 17.
The processor 41 may be a Central Processing Unit (CPU). The Processor 41 may also be other general purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs) or other Programmable logic devices, discrete Gate or transistor logic devices, discrete hardware components, or combinations thereof. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The memory 42 may include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created from use of the travel target point recognition model evaluation device, or the travel target point recognition model development device, or the travel target point recognition device, or the like. Further, the memory 42 may include high speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, the memory 42 may optionally include a memory remotely disposed with respect to the processor 41, and these remote memories may be connected to the travel target point recognition model evaluation device, or the travel target point recognition model development device, or the travel target point recognition device through a network. The input device 43 may receive a calculation request (or other numerical or character information) input by a user and generate a key signal input in relation to the travel target point recognition model evaluation device, or the travel target point recognition model development device, or the travel target point recognition device. The output device 44 may include a display device such as a display screen for outputting the calculation result.
Example 8
The present invention provides a computer-readable storage medium, which stores computer instructions, and it can be understood by those skilled in the art that all or part of the processes in the methods of the above embodiments can be implemented by a computer program to instruct related hardware, where the program can be stored in a computer-readable storage medium, and when executed, the program can include the processes of the embodiments of the methods. The storage medium may be a magnetic Disk, an optical Disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a Flash Memory (Flash Memory), a Hard Disk (Hard Disk Drive, abbreviated as HDD), a Solid State Drive (SSD), or the like; the storage medium may also comprise a combination of memories of the kind described above.
It should be understood that the above examples are only for clarity of illustration and are not intended to limit the embodiments. Other variations and modifications will be apparent to persons skilled in the art in light of the above description. And are neither required nor exhaustive of all embodiments. And obvious variations or modifications therefrom are within the scope of the invention.

Claims (12)

1. A travel target point identification model evaluation method is characterized by comprising the following steps:
identifying the travel data according to a preset travel target point identification model to generate an identification conversion point sequence;
acquiring a real conversion point sequence of the travel data;
calculating a preset model evaluation index according to the identification conversion point sequence and the real conversion point sequence, wherein the preset model evaluation index comprises a target behavior start-stop moment error, a target behavior duration error, a target behavior center point offset distance and accuracy;
and determining the evaluation result of the travel target point identification model according to the preset model evaluation index.
2. The method for evaluating a travel target point identification model according to claim 1, wherein the step of calculating a preset model evaluation index according to the identification transition point sequence and the real transition point sequence comprises:
acquiring the recognition starting time and the recognition ending time of the target behavior according to the recognition transition point sequence;
acquiring the real starting time and the real ending time of the target behavior according to the real conversion point sequence;
and determining the target behavior start-stop time error according to the identification start time, the identification stop time, the real start time and the real stop time.
3. The method for evaluating a travel target point identification model according to claim 2, wherein the step of calculating a preset model evaluation index according to the identification transition point sequence and the real transition point sequence further comprises:
calculating the recognition duration and the actual duration of the target behavior according to the recognition starting time, the recognition ending time, the real starting time and the real ending time;
and determining the target behavior duration error according to the identification duration and the actual duration.
4. The method for evaluating a travel target point identification model according to claim 1, wherein the step of calculating a preset model evaluation index according to the identification transition point sequence and the real transition point sequence comprises:
acquiring the latitude and longitude of the central point of the identification conversion point sequence according to the identification conversion point sequence;
acquiring the latitude and longitude of the central point of the real conversion point sequence according to the real conversion point sequence;
and determining the target behavior center point offset distance according to the radius of the earth, the latitude and the longitude of the center point of the identification conversion point sequence and the latitude and the longitude of the center point of the real conversion point sequence.
5. The method for evaluating the travel target point identification model according to any one of claims 1 to 4, wherein the step of determining the evaluation result of the travel target point identification model according to the preset model evaluation index includes:
and if the target behavior start-stop moment error, the target behavior duration error, the target behavior center point offset distance and the accuracy are all larger than respective threshold values, judging that the preset travel target point identification model is qualified.
6. A travel target point identification model development method is characterized by comprising the following steps:
acquiring a plurality of pieces of travel data, wherein the travel data comprise a training set and a verification set;
extracting the characteristic value of each sampling point in the training set and the characteristic value of each sampling point in the verification set;
developing an initial travel target point identification model according to a preset integrated learning method and the characteristic values of all sampling points in the training set;
acquiring a real conversion point sequence according to the verification set;
inputting the characteristic values of the sampling points in the verification set into the initial trip target point identification model to obtain an identification conversion point sequence;
calculating a preset model evaluation index according to the identification conversion point sequence and the real conversion point sequence, wherein the preset model evaluation index comprises a target behavior start-stop moment error, a target behavior duration error, a target behavior center point offset distance and accuracy;
determining an evaluation result of the initial trip target point identification model according to the preset model evaluation index;
if the evaluation result is qualified, determining the initial trip target point identification model as the trip target point identification model;
and if the evaluation result is unqualified, returning to the step of extracting the characteristic value of each sampling point in the training set and the characteristic value of each sampling point in the verification set, or developing an initial trip target point identification model according to a preset integrated learning method and the characteristic values of each sampling point in the training set.
7. A travel target point identification method is characterized by comprising the following steps:
acquiring outgoing data to be predicted;
extracting characteristic values of sampling points in the to-be-predicted row data;
inputting the characteristic values of the sampling points in the outgoing data to be predicted into an outgoing target point identification model to obtain a target point sequence, wherein the outgoing target point identification model is obtained according to the outgoing target point identification model development method of claim 6.
8. A travel target point recognition model evaluation device is characterized by comprising:
the system comprises a travel target point identification module, a travel conversion point sequence acquisition module and a travel conversion point sequence generation module, wherein the travel target point identification module is used for identifying travel data according to a preset travel target point identification model and generating an identification conversion point sequence;
the real conversion point sequence acquisition module is used for acquiring a real conversion point sequence of the trip data;
the preset model evaluation index calculation module is used for calculating preset model evaluation indexes according to the identification transition point sequence and the real transition point sequence, wherein the preset model evaluation indexes comprise target behavior starting and stopping time errors, target behavior duration errors, target behavior center point offset distances and accuracy rates;
and the travel target point identification model evaluation module is used for determining the evaluation result of the travel target point identification model according to the preset model evaluation index.
9. A travel target point identification model development device is characterized by comprising:
the trip data acquisition module is used for acquiring a plurality of trip data, and the trip data comprises a training set and a verification set;
the characteristic value extraction module is used for extracting the characteristic values of the sampling points in the training set and the characteristic values of the sampling points in the verification set;
the initial travel target point identification model establishing module is used for developing an initial travel target point identification model according to a preset integrated learning method and the characteristic values of all sampling points in the training set;
the real conversion point sequence acquisition module is used for acquiring a real conversion point sequence according to the verification set;
the identification conversion point sequence acquisition module is used for inputting the characteristic values of the sampling points in the verification set into the initial trip target point identification model to acquire an identification conversion point sequence;
the preset model evaluation index calculation module is used for calculating preset model evaluation indexes according to the identification transition point sequence and the real transition point sequence, wherein the preset model evaluation indexes comprise target behavior starting and stopping time errors, target behavior duration errors, target behavior center point offset distances and accuracy rates;
the initial travel target point identification model evaluation module is used for determining an evaluation result of the initial travel target point identification model according to the preset model evaluation index;
and the trip target point identification model judging module is used for judging the initial trip target point identification model, if the evaluation result is qualified, the initial trip target point identification model is determined as the trip target point identification model, and if the evaluation result is unqualified, the characteristic value extracting module is triggered to execute, or the initial trip target point identification model establishing module is triggered to execute.
10. A travel target point identifying apparatus, comprising:
the device comprises a to-be-predicted outgoing data acquisition module, a to-be-predicted outgoing data acquisition module and a to-be-predicted outgoing data acquisition module, wherein the to-be-predicted outgoing data acquisition module is used for acquiring outgoing data to be predicted;
the system comprises a to-be-predicted row data characteristic value extraction module, a to-be-predicted row data characteristic value extraction module and a to-be-predicted row data characteristic value extraction module, wherein the to-be-predicted row data characteristic value extraction module is used for extracting characteristic values of all sampling points in the to-be-predicted row data;
a target point sequence identification module, configured to input the feature values of the sampling points in the outgoing data to be predicted into an outgoing target point identification model, and obtain a target point sequence, where the outgoing target point identification model is obtained according to the outgoing target point identification model development method according to claim 6.
11. A computer device, comprising:
at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to perform the travel target point recognition model evaluation method of any one of claims 1-5, or the travel target point recognition model development method of claim 6, or the travel target point recognition method of claim 7.
12. A computer-readable storage medium storing computer instructions for causing a computer to execute the travel target point identification model evaluation method according to any one of claims 1 to 5, or the travel target point identification model development method according to claim 6, or the travel target point identification method according to claim 7.
CN201910925488.1A 2019-09-27 2019-09-27 Travel target point identification method and device, and model development and evaluation method and device Pending CN110675074A (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201910994735.3A CN110728459A (en) 2019-09-27 2019-09-27 Travel mode identification system, method and device and model training method and device
CN201910925488.1A CN110675074A (en) 2019-09-27 2019-09-27 Travel target point identification method and device, and model development and evaluation method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910925488.1A CN110675074A (en) 2019-09-27 2019-09-27 Travel target point identification method and device, and model development and evaluation method and device

Related Child Applications (1)

Application Number Title Priority Date Filing Date
CN201910994735.3A Division CN110728459A (en) 2019-09-27 2019-09-27 Travel mode identification system, method and device and model training method and device

Publications (1)

Publication Number Publication Date
CN110675074A true CN110675074A (en) 2020-01-10

Family

ID=69079581

Family Applications (2)

Application Number Title Priority Date Filing Date
CN201910925488.1A Pending CN110675074A (en) 2019-09-27 2019-09-27 Travel target point identification method and device, and model development and evaluation method and device
CN201910994735.3A Pending CN110728459A (en) 2019-09-27 2019-09-27 Travel mode identification system, method and device and model training method and device

Family Applications After (1)

Application Number Title Priority Date Filing Date
CN201910994735.3A Pending CN110728459A (en) 2019-09-27 2019-09-27 Travel mode identification system, method and device and model training method and device

Country Status (1)

Country Link
CN (2) CN110675074A (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111882102B (en) * 2020-06-05 2024-03-22 百度在线网络技术（北京）有限公司 Compensation method, equipment and storage medium for interruption interval information in track
CN114757451B (en) * 2022-06-14 2022-10-25 北京航空航天大学 Dynamic traffic trip demand prediction method for large activities
CN116088321B (en) * 2023-04-12 2023-07-04 宁波吉利汽车研究开发有限公司 Automatic driving decision control method and device and electronic equipment

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102682654A (en) * 2011-03-16 2012-09-19 高德软件有限公司 Method and device for rendering traffic information
CN105547306A (en) * 2015-08-11 2016-05-04 深圳大学 Route pushing method and system thereof
CN106931974A (en) * 2017-03-29 2017-07-07 清华大学 The method that personal Commuting Distance is calculated based on mobile terminal GPS location data record
CN108775900A (en) * 2018-07-31 2018-11-09 上海哔哩哔哩科技有限公司 Phonetic navigation method, system based on WEB and storage medium
CN110276563A (en) * 2019-07-01 2019-09-24 长安大学 A kind of mode of transportation transfer Activity recognition method based on supporting vector machine model

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102682654A (en) * 2011-03-16 2012-09-19 高德软件有限公司 Method and device for rendering traffic information
CN105547306A (en) * 2015-08-11 2016-05-04 深圳大学 Route pushing method and system thereof
CN106931974A (en) * 2017-03-29 2017-07-07 清华大学 The method that personal Commuting Distance is calculated based on mobile terminal GPS location data record
CN108775900A (en) * 2018-07-31 2018-11-09 上海哔哩哔哩科技有限公司 Phonetic navigation method, system based on WEB and storage medium
CN110276563A (en) * 2019-07-01 2019-09-24 长安大学 A kind of mode of transportation transfer Activity recognition method based on supporting vector machine model

Also Published As

Publication number Publication date
CN110728459A (en) 2020-01-24

Similar Documents

Publication Publication Date Title
CN110675074A (en) Travel target point identification method and device, and model development and evaluation method and device
CN108846063B (en) Method, device, equipment and computer readable medium for determining answers to questions
US20220383627A1 (en) Automatic modeling method and device for object detection model
CN110262273A (en) A kind of home equipment control method, device, storage medium and smart home system
CN106503231B (en) Search method and device based on artificial intelligence
CN110489521B (en) Text type detection method and device, electronic equipment and computer readable medium
WO2019080484A1 (en) Method of pruning convolutional neural network based on feature map variation
CN111382255A (en) Method, apparatus, device and medium for question and answer processing
CN104021674A (en) Method for rapidly and accurately forecasting travel time of vehicles for passing through road sections
US11514315B2 (en) Deep neural network training method and apparatus, and computer device
CN113256326A (en) Method for realizing prediction of commodity extra-large screen point position pedestrian volume based on deep learning
CN112100374A (en) Text clustering method and device, electronic equipment and storage medium
CN113361690A (en) Water quality prediction model training method, water quality prediction device, water quality prediction equipment and medium
CN111144462A (en) Unknown individual identification method and device for radar signals
CN114154622A (en) Algorithm model for traffic operation system flow data acquisition missing completion
JP2020067910A (en) Learning curve prediction device, learning curve prediction method, and program
CN112733552A (en) Machine translation model construction method, device and equipment
CN111160594B (en) Method and device for estimating arrival time and storage medium
CN112148819A (en) Address recognition method and device combining RPA and AI
CN111325255A (en) Specific crowd delineating method and device, electronic equipment and storage medium
CN114245329B (en) Traffic mode identification method, device, equipment and storage medium
CN113792132B (en) Target answer determining method, device, equipment and medium
CN115391523A (en) Wind power plant multi-source heterogeneous data processing method and device
CN115222018A (en) Non-invasive load decomposition method and device
CN114911814A (en) Consultation service method and system based on knowledge resource library updating

Legal Events

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

Application publication date: 20200110

RJ01 Rejection of invention patent application after publication