CN111832769A

CN111832769A - Method and device for ordering vehicle-entering points and information

Info

Publication number: CN111832769A
Application number: CN201910906430.2A
Authority: CN
Inventors: 陈欢; 陶言祺
Original assignee: Beijing Didi Infinity Technology and Development Co Ltd
Current assignee: Beijing Didi Infinity Technology and Development Co Ltd
Priority date: 2019-09-24
Filing date: 2019-09-24
Publication date: 2020-10-27

Abstract

The application provides a method and a device for ordering vehicle-entering points and information, wherein the method comprises the following steps: acquiring geographic position information of a service request end, and determining a plurality of alternative boarding points based on the geographic position information; for each alternative boarding point, inputting a target characteristic value of the alternative boarding point under a plurality of correlation characteristics into a pre-trained correlation detection model, and acquiring the correlation between the alternative boarding point and a service request end; sorting the alternative vehicle-entering points according to the correlation; the relevancy detection model comprises a plurality of relevancy decision trees; and training a plurality of correlation decision trees and a plurality of noise decision trees alternately, wherein the residual error of any one correlation decision tree is the fitting target of the next noise decision tree with the closest distance to the any one correlation decision tree, and the residual error of any one noise decision tree is the fitting target of the next correlation decision tree with the closest distance to the any one noise decision tree. The accuracy of getting on bus point sequencing can be promoted in the embodiment of the application.

Description

Method and device for ordering vehicle-entering points and information

Technical Field

The application relates to the technical field of computer application, in particular to a method and a device for ordering boarding points and ordering information.

Background

With the continuous development of network technology and mobile terminal technology, people increasingly rely on mobile terminals to obtain information; with the explosive growth of data, more and more information can be obtained by people through mobile terminals; however, the amount of information that can be displayed by the mobile terminal for the user is limited, and the obtained information cannot be displayed completely, which involves a process of sorting the obtained information.

Taking the field of online car booking as an example, users of the online car booking platform comprise a service requester and a service provider. The service request party initiates a car booking request to a server of the network car booking platform based on the service request terminal; the server matches a plurality of service providing terminals for the service request party based on a certain algorithm, and forwards the car booking request to each matched service providing terminal. In the process that a user request party initiates a taxi appointment request to a server based on a service request terminal, the service request terminal determines a plurality of alternative taxi-entering points for the service request party, sorts the alternative taxi-entering points based on a certain mode, and then displays a certain number of the alternative taxi-entering points to the service request party according to the sorting for the service request party to select. The current method for sequencing all the alternative boarding points is generally carried out according to the positions of all the alternative boarding points and a service request end; the ordering mode causes that the boarding point determined for the service requester is not the optimal boarding position, for example, the problems of complex route, difficult finding and the like exist.

The current information sorting mode has the problem of low sorting accuracy.

Disclosure of Invention

In view of the above, an object of the present application is to provide a method and an apparatus for vehicle-entering point ranking and information ranking, which can train a correlation detection model through correlation characteristics and noise characteristics; based on the relevance detection model obtained in the mode, the relevance between the alternative information and the user can be accurately obtained; and further, higher sorting accuracy is achieved when sorting is performed based on the relevance and the alternative information.

In a first aspect, an embodiment of the present application provides a method for ordering boarding points, including:

acquiring geographic position information of a service request end, and determining a plurality of alternative boarding points based on the geographic position information; for each alternative boarding point, inputting a target characteristic value of the alternative boarding point under a plurality of correlation characteristics into a pre-trained correlation detection model, and acquiring the correlation between the alternative boarding point and the service request end; sequencing the alternative boarding points according to the corresponding correlation degree of each alternative boarding point so that a user can select a final boarding point according to a sequencing result; wherein the relevance detection model comprises a plurality of relevance decision trees; and training the plurality of the correlation decision trees and the plurality of noise decision trees alternately, wherein during training, the residual error of any one correlation decision tree is the fitting target of the next noise decision tree with the closest distance to the any one correlation decision tree, and the residual error of any one noise decision tree is the fitting target of the next correlation decision tree with the closest distance to the any one noise decision tree.

In an alternative embodiment, the correlation detection model is trained in the following way:

determining a plurality of the sample loading points; obtaining a first sample characteristic value of each sample getting-on point under a plurality of correlation characteristics and a second sample characteristic value under a plurality of noise characteristics; constructing a plurality of correlation decision trees corresponding to the correlation features based on the first sample feature values, and constructing a plurality of noise decision trees corresponding to the noise features based on the second sample feature values; and taking the constructed multiple relevance decision trees as the relevance detection model.

In an alternative embodiment, the sample loading station comprises: a positive sample loading point and a negative sample loading point; the determining the plurality of sample getting-on points specifically includes: obtaining sample order information of a plurality of sample orders; the sample order information includes: a sample display boarding point and a sample selection boarding point; determining the selected getting-on point of the sample as a getting-on point of a positive sample; and determining other vehicle-loading points except the selected vehicle-loading point of the sample in the sample display vehicle-loading points as negative sample vehicle-loading points.

In an optional implementation manner, the obtaining geographic location information of the service request end and determining a plurality of candidate boarding points based on the geographic location information includes: receiving the geographical position information sent by a service request end when a target service interface is opened; based on the geographic position information, obtaining a plurality of alternative boarding points corresponding to the geographic position information from a boarding point database; and the distance between the geographical position indicated by the geographical position information and the alternative vehicle-entering point is smaller than a preset distance threshold value.

In an alternative embodiment, the correlation features include one or more of: the distance between the alternative vehicle getting-on point and the service request end, whether the alternative vehicle getting-on point and the service request end are located on the same side of the same road section, the road grade of the road section where the alternative vehicle getting-on point is located, the first heat value of the road section where the alternative vehicle getting-on point is used as the selected vehicle getting-on point, the second heat value of the road section where the alternative vehicle getting-on point is located, and the third heat value of the alternative vehicle getting-on point selected after being displayed to the user.

In an optional implementation manner, for a case that the correlation feature includes a distance between the candidate boarding point and a service request end, the distance between the candidate boarding point and the service request end is obtained in the following manner: obtaining the distance between the alternative boarding point and the service request end according to the geographic position coordinate corresponding to the alternative boarding point and the geographic position information of the service request end; for the case that the correlation feature includes the first heat value, the first heat value is obtained in the following manner: acquiring order information of a historical order; the order information comprises selected boarding point information; counting a first quantity of historical orders taking the alternative boarding points as the selected boarding points based on the selected boarding point information, and taking the first quantity as the first heat value;

for the case that the correlation feature includes the second heat value, the second heat value is obtained in the following manner: obtaining order information of a historical order and determining the vehicle-entering points on the same road section of the same road section with the alternative vehicle-entering points; the order information comprises selected boarding point information; according to the fact that the order information comprises selected boarding points information, counting a first quantity of historical orders taking the alternative boarding points as the selected boarding points, and counting a second quantity of the historical orders taking the boarding points on the same road section as the selected boarding points; taking the sum of the first amount and the second amount as the second heat value;

for a case that the correlation feature includes the third heat value, obtaining the third heat value in the following manner: acquiring order information of a historical order; the order information comprises: displaying the information of the boarding points and selecting the information of the boarding points; according to the displayed pick-up point information and the selected pick-up point information, counting a third quantity of historical orders when the alternative pick-up points are used as the selected pick-up points; taking the third amount as the third heat value.

In an alternative embodiment, the noise signature includes one or more of: and when the sample boarding point is taken as the selected boarding point, the instant call duration, the driving receiving distance and the driving receiving time between the service request end and the service providing end are determined.

In an optional implementation manner, after the sorting the candidate pick-up points according to the respective corresponding relevancy of each candidate pick-up point, the method further includes: and determining a preset number of fixed point boarding points from the alternative boarding points according to the sequence of the alternative boarding points, and displaying the fixed point boarding points through the service request end.

In a second aspect, an embodiment of the present application provides an information sorting method, including:

acquiring user information, and determining a plurality of pieces of alternative information corresponding to the user information based on the user information; for each piece of alternative information, inputting the target characteristic value of the alternative information under a plurality of correlation characteristics into a pre-trained correlation detection model, and acquiring the correlation between the alternative information and the user information; sorting the alternative information according to the corresponding relevancy of each alternative information so that a user can select final alternative information according to a sorting result; wherein the relevance detection model comprises a plurality of relevance decision trees; and training the plurality of the correlation decision trees and the plurality of noise decision trees alternately, wherein during training, the residual error of any one correlation decision tree is the fitting target of the next noise decision tree with the closest distance to the any one correlation decision tree, and the residual error of any one noise decision tree is the fitting target of the next correlation decision tree with the closest distance to the any one noise decision tree.

determining a plurality of pieces of the sample information; acquiring a first sample characteristic value of each piece of sample information under a plurality of correlation characteristics and a second sample characteristic value under a plurality of noise characteristics; constructing a plurality of correlation decision trees corresponding to the correlation features based on the first sample feature values, and constructing a plurality of noise decision trees corresponding to the noise features based on the second sample feature values; and taking the constructed multiple relevance decision trees as the relevance detection model.

In a third aspect, an embodiment of the present application provides an apparatus for ordering boarding points, including:

the system comprises a first acquisition module, a second acquisition module and a third acquisition module, wherein the first acquisition module is used for acquiring the geographic position information of a service request end and determining a plurality of alternative boarding points based on the geographic position information;

the first determining module is used for inputting the target characteristic values of the candidate boarding points under the multiple correlation characteristics into a pre-trained correlation detection model aiming at each candidate boarding point, and acquiring the correlation between the candidate boarding points and the service request end;

the first sequencing module is used for sequencing the alternative boarding points according to the corresponding relevancy of each alternative boarding point so that a user can select a final boarding point according to a sequencing result;

wherein the relevance detection model comprises a plurality of relevance decision trees; and training the plurality of the correlation decision trees and the plurality of noise decision trees alternately, wherein during training, the residual error of any one correlation decision tree is the fitting target of the next noise decision tree with the closest distance to the any one correlation decision tree, and the residual error of any one noise decision tree is the fitting target of the next correlation decision tree with the closest distance to the any one noise decision tree.

In an alternative embodiment, the method further comprises: a first training module for training the correlation detection model in the following manner:

In an alternative embodiment, the sample loading station comprises: a positive sample loading point and a negative sample loading point;

the first training module is specifically configured to determine a plurality of the sample boarding points in the following manner:

obtaining sample order information of a plurality of sample orders; the sample order information includes: a sample display boarding point and a sample selection boarding point;

determining the selected getting-on point of the sample as a getting-on point of a positive sample;

and determining other vehicle-loading points except the selected vehicle-loading point of the sample in the sample display vehicle-loading points as negative sample vehicle-loading points.

In an optional implementation manner, the first obtaining module is configured to obtain geographic location information of a service request end and determine a plurality of candidate boarding points based on the geographic location information, where: receiving the geographical position information sent by a service request end when a target service interface is opened; based on the geographic position information, obtaining a plurality of alternative boarding points corresponding to the geographic position information from a boarding point database; and the distance between the geographical position indicated by the geographical position information and the alternative vehicle-entering point is smaller than a preset distance threshold value.

In an optional implementation manner, the first obtaining module is further configured to:

aiming at the condition that the correlation characteristics comprise the distance between the candidate getting-on point and the service request end, the distance between the candidate getting-on point and the service request end is obtained in the following mode: obtaining the distance between the alternative boarding point and the service request end according to the geographic position coordinate corresponding to the alternative boarding point and the geographic position information of the service request end;

for the case that the correlation feature includes the first heat value, the first heat value is obtained in the following manner: acquiring order information of a historical order; the order information comprises selected boarding point information; counting a first quantity of historical orders taking the alternative boarding points as the selected boarding points based on the selected boarding point information, and taking the first quantity as the first heat value;

In an alternative embodiment, the method further comprises: and the first display module is used for determining a preset number of fixed point boarding points from the alternative boarding points according to the sequence of the alternative boarding points and displaying the fixed point boarding points through the service request end.

In a fourth aspect, an embodiment of the present application further provides an information sorting apparatus, including:

the second acquisition module is used for acquiring user information and determining a plurality of pieces of alternative information corresponding to the user information based on the user information;

the second determining module is used for inputting the target characteristic values of each piece of candidate information under a plurality of correlation characteristics into a pre-trained correlation degree detection model and acquiring the correlation degree between the candidate information and the user information;

the second sorting module is used for sorting the alternative information according to the corresponding relevancy of each alternative information so that a user can select the final alternative information according to a sorting result;

In an alternative embodiment, the method further comprises: a second training module for training the correlation detection model in the following manner:

In a fifth aspect, an embodiment of the present application further provides a computer device, including: a processor, a memory and a bus, the memory storing machine-readable instructions executable by the processor, the processor and the memory communicating via the bus when the computer device is running, the machine-readable instructions when executed by the processor performing the steps of the first aspect, or any of the possible implementations of the first aspect, or the second aspect, or any of the possible implementations of the second aspect.

In a fourth aspect, this embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, performs the steps in the first aspect, or any one of the possible implementations of the first aspect, or the second aspect, or any one of the possible implementations of the second aspect.

According to the method and the device, a plurality of candidate getting-on points are obtained through geographic position information of a service request end, target characteristic values of the candidate getting-on points under a plurality of correlation characteristics are input into a pre-trained correlation degree detection model aiming at each candidate getting-on point, correlation degrees between each candidate getting-on point and the service request end are obtained, and the candidate getting-on points are sorted based on the correlation degrees corresponding to the candidate getting-on points; the relevancy detection model comprises a plurality of relevancy decision trees; the plurality of correlation decision trees and the plurality of noise decision trees are alternately trained, in the training process, the residual error of any one correlation decision tree is the fitting target of the next noise decision tree closest to the any one correlation decision tree, the residual error of any one noise decision tree is the fitting target of the next correlation decision tree closest to the any one noise decision tree, in the using process, only the correlation decision tree is used as a correlation detection model to obtain the correlation of each alternative boarding point, further, in the process of training the correlation detection model, the influence of correlation characteristics and noise characteristics on the model is considered, in the using process, the influence of the noise characteristics on correlation detection results is eliminated, and the higher correlation detection precision is achieved, and further, when the boarding points are sequenced, the higher sequencing accuracy is achieved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are required to be used in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present application and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained from the drawings without inventive effort.

Fig. 1 is a schematic diagram illustrating an architecture of a service system provided in an embodiment of the present application;

FIG. 2 is a flow chart illustrating a method for pick-up point ranking provided by an embodiment of the present application;

fig. 3 is a flowchart illustrating a specific method for training a correlation detection model in the method for sorting vehicle-entering points provided by the embodiment of the present application;

FIG. 4 is a schematic structural diagram illustrating an apparatus for entering a train point sequence provided by an embodiment of the present application;

fig. 5 shows a schematic structural diagram of a computer device 50 provided by an embodiment of the present application;

FIG. 6 is a flow chart illustrating a method for ordering information provided by an embodiment of the present application;

FIG. 7 is a schematic structural diagram illustrating an apparatus for sorting information according to an embodiment of the present disclosure;

fig. 8 shows a schematic structural diagram of a computer device 80 according to an embodiment of the present application.

Detailed Description

In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it should be understood that the drawings in the present application are for illustrative and descriptive purposes only and are not used to limit the scope of protection of the present application. Additionally, it should be understood that the schematic drawings are not necessarily drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of the present application. It should be understood that the operations of the flow diagrams may be performed out of order, and steps without logical context may be performed in reverse order or simultaneously. One skilled in the art, under the guidance of this application, may add one or more other operations to, or remove one or more operations from, the flowchart.

In addition, the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. The components of the embodiments of the present application, generally described and illustrated in the figures herein, can be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of the present application, presented in the accompanying drawings, is not intended to limit the scope of the claimed application, but is merely representative of selected embodiments of the application. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present application without making any creative effort, shall fall within the protection scope of the present application.

To enable those skilled in the art to use the present disclosure, the following embodiments are presented in conjunction with a specific application scenario, "pick-up ordering". It will be apparent to those skilled in the art that the general principles defined herein may be applied to other embodiments and applications without departing from the spirit and scope of the application. Although the present application is primarily described with respect to a plurality of alternative pick-up point orderings determined for a user based on a current location of a service requester, it should be understood that this is merely one exemplary embodiment. Other information may also be ranked, for example, a plurality of candidate points of interest (POIs) retrieved through POI retrieval information may be ranked, a plurality of advertisements to be pushed to the user may be ranked, and the like.

It should be noted that in the embodiments of the present application, the term "comprising" is used to indicate the presence of the features stated hereinafter, but does not exclude the addition of further features.

The terms "passenger," "requestor," "service requestor," and "customer" are used interchangeably in this application to refer to an individual, entity, or tool that can request or order a service. The terms "driver," "provider," "service provider," and "provider" are used interchangeably in this application to refer to an individual, entity, or tool that can provide a service. The term "user" in this application may refer to an individual, entity or tool that requests a service, subscribes to a service, provides a service, or facilitates the provision of a service. For example, the user may be a passenger, a driver, an operator, etc., or any combination thereof. In the present application, "passenger" and "passenger terminal" may be used interchangeably, and "driver" and "driver terminal" may be used interchangeably.

The terms "service request" and "order" are used interchangeably herein to refer to a request initiated by a passenger, a service requester, a driver, a service provider, or a supplier, the like, or any combination thereof. Accepting the "service request" or "order" may be a passenger, a service requester, a driver, a service provider, a supplier, or the like, or any combination thereof. The service request may be charged or free.

The Positioning technology used in the present application may be based on a Global Positioning System (GPS), a Global Navigation Satellite System (GLONASS), a COMPASS Navigation System (COMPASS), a galileo Positioning System, a Quasi-Zenith Satellite System (QZSS), a Wireless Fidelity (WiFi) Positioning technology, or the like, or any combination thereof. One or more of the above-described positioning systems may be used interchangeably in this application.

One aspect of the present application relates to a system for pick-up ordering. The system can acquire a plurality of alternative vehicle-entering points through the geographical position information of the service request end, and inputs target characteristic values of the alternative vehicle-entering points under a plurality of correlation characteristics into a pre-trained correlation degree detection model aiming at each alternative vehicle-entering point, acquires the correlation degree between each alternative vehicle-entering point and the service request end, and sorts the alternative vehicle-entering points based on the correlation degree corresponding to each alternative vehicle-entering point; the relevancy detection model comprises a plurality of relevancy decision trees; the method comprises the steps that a plurality of correlation decision trees and a plurality of noise decision trees are alternately trained, in the training process, the residual error of any one correlation decision tree is the fitting target of the next noise decision tree closest to the correlation decision tree, the residual error of any one noise decision tree is the fitting target of the next correlation decision tree closest to the noise decision tree, in the using process, only the correlation decision tree is used as a correlation detection model to obtain the correlation of each alternative boarding point, further, in the process of training the correlation detection model, the influence of correlation characteristics and noise characteristics on the model is considered, in the using process, the influence of the noise characteristics on correlation detection results is eliminated, the higher correlation detection precision is achieved, and further, when the boarding points are sequenced, the higher sequencing accuracy is achieved.

Fig. 1 is a schematic architecture diagram of a service system 100 for implementing pick-up ordering according to an embodiment of the present application. For example, the service system 100 may be an online transportation service platform for transportation services such as taxi cab, designated drive service, express, carpool, bus service, driver rental, or shift service, or any combination thereof. The service system 100 may include one or more of a server 110, a network 120, a service requester 130, a service provider 140, and a database 150.

In some embodiments, the server 110 may include a processor. The processor may process information and/or data related to the service request to perform one or more of the functions described herein. For example, the processor may obtain the candidate pick-up points based on the service request obtained from the service request terminal 130, sort the candidate pick-up points, and feed back a preset number of pick-up points to be displayed to the service request terminal 130 according to the sort of each candidate pick-up point. In some embodiments, a processor may include one or more processing cores (e.g., a single-core processor (S) or a multi-core processor (S)). Merely by way of example, a Processor may include a Central Processing Unit (CPU), an Application Specific Integrated Circuit (ASIC), an Application specific Instruction Set Processor (ASIP), a Graphics Processing Unit (GPU), a Physical Processing Unit (PPU), a Digital Signal Processor (DSP), a Field Programmable Gate Array (FPGA), a Programmable Logic Device (PLD), a controller, a microcontroller Unit, a Reduced Instruction Set computer (Reduced Instruction Set Computing, RISC), a microprocessor, or the like, or any combination thereof.

In some embodiments, the device types corresponding to the service request end 130 and the service providing end 140 may be mobile devices, such as smart home devices, wearable devices, smart mobile devices, virtual reality devices, or augmented reality devices, and the like, and may also be tablet computers, laptop computers, or built-in devices in motor vehicles, and the like.

In some embodiments, a database 150 may be connected to the network 120 to communicate with one or more components (e.g., the server 110, the service requester 130, the service provider 140, etc.) in the service system 100. One or more components in the service system 100 may access data or instructions stored in the database 150 via the network 120. In some embodiments, the database 150 may be directly connected to one or more components in the service system 100, or the database 150 may be part of the server 110.

The following describes the information ranking method provided in the embodiment of the present application in detail with reference to the content described in the service system 100 shown in fig. 1.

Referring to fig. 2, a schematic flow chart of a method for sorting pick-up points according to an embodiment of the present application is shown, where the method may be executed by the server 110 in the service system 100, or may be executed by the service request end 130, and the specific execution process is as follows:

s201: and acquiring the geographical position information of the service request terminal, and determining a plurality of alternative boarding points based on the geographical position information.

S202: and for each candidate getting-on point, inputting the target characteristic value of the candidate getting-on point under a plurality of correlation characteristics into a pre-trained correlation detection model, and acquiring the correlation between the candidate getting-on point and the service request end.

S203: and sequencing the alternative vehicle-entering points according to the corresponding correlation degree of each alternative vehicle-entering point.

The following is a description of the above S201 to S203.

I: in the above S201, the service request end displays a service page capable of providing different services for the user. For example: a car appointment service page, an invoice service page, a trip inquiry service page, a customer service page, etc. One or more of which may be set as a target service page. When the target service page is opened, the service request end sends the geographical position information of the service request end to the server.

Exemplarily, the car appointment service page is taken as a target service page; when a user opens the closed service software on the service request end and enters the taxi appointment service page, the service request end acquires the geographical position information of the user and sends the geographical position information to the server; when a user controls service software running on a background to be switched to a foreground to run on a service request end and enters the taxi appointment service page, the service request end automatically acquires geographical position information of the service software and sends the geographical position information to a server; and when the service request end transfers from other service pages to the taxi appointment service page, the service request end automatically acquires the geographical position information of the service request end and sends the geographical position information to the server.

It should be noted that, in order to reduce the traffic consumption and the waste of server computing resources caused by frequent sending of geographical location information to the server due to frequent opening of the target service page, the service request end may detect a time difference between two adjacent times of entering the target service page; if the time difference is smaller than the preset time difference threshold, the service request end may not resend the geographic location information to the server, but redisplay the content of the target service page previously displayed for the user.

In practical application, if the execution main body of the method for ordering the boarding points is the service request end, the service request end sends the geographical position information to the server after acquiring the geographical position information of the service request end. After receiving the geographic position information, the server acquires a plurality of alternative boarding points corresponding to the geographic position information from the boarding point database based on the geographic position information, and then sends the acquired relevant information of the alternative boarding points to the service request terminal so that the service request terminal can sequence the alternative boarding points. If the execution main body of the boarding point sequencing method is the server, after receiving the geographic position information sent by the service request end, the server can acquire a plurality of alternative boarding points corresponding to the geographic position information from the boarding point database, and then sequence each alternative boarding point.

Here, the boarding point database stores a plurality of preset boarding points; each boarding point corresponds to a boarding point name and a geographical position coordinate corresponding to the boarding point; the name of the boarding point is used for describing the specific position of the boarding point in text, such as 'east door of XX cell', 'west-south side of XX crossing', and the like, and the geographic position coordinate is used for identifying the specific geographic position of the boarding point.

When the alternative boarding points are obtained from the boarding point database, the distance between the boarding points in the database and the server request end can be calculated according to the geographical position coordinates of each boarding point in the boarding point database and the geographical position information of the service request end, and the boarding points with the distance smaller than the preset distance threshold value are determined as the alternative boarding points.

In addition, the vehicle-entering points in the database can be grouped according to the geographic position coordinates of the vehicle-entering points to form a plurality of vehicle-entering point sets, and the central geographic position coordinates are determined for each vehicle-entering point set; when the alternative boarding points are obtained, the distance between the service request end and the position indicated by each central geographic position coordinate can be calculated according to the geographic position information of the service request end and the central geographic position coordinate corresponding to each boarding point set; then, the boarding points corresponding to the central geographic position coordinates with the distance smaller than the preset distance are combined into a target boarding point set, and candidate boarding points with the distance smaller than a preset distance threshold value with the geographic position indicated by the geographic position information of the service request end are screened from the target boarding point set, so that the calculation amount in the process of determining the candidate boarding points can be reduced.

It should be noted here that the target boarding point set corresponding to the service request side may be one or more.

II: in S202, before inputting the target feature values of the candidate boarding points under the multiple correlation features into the pre-trained correlation detection model and obtaining the correlation between the candidate boarding points and the service request end, the target feature values of the candidate boarding points under the multiple correlation features are obtained first. The correlation feature is a feature that can have a positive influence on a parameter of the model, and is a prior feature, that is, a feature value can be determined for each candidate boarding point before the boarding point is determined.

Specifically, the relevant characteristics include, but are not limited to, one or more of the following a 1-a 6:

a 1: the distance between the alternative boarding point and the service request end.

Here, each boarding point corresponds to a boarding point name and a geographical position coordinate corresponding to the boarding point. When the alternative boarding point is obtained, the geographic position coordinate and the boarding point name of the alternative boarding point can be obtained, and then the distance between the alternative boarding point and the service request end can be calculated according to the geographic position coordinate of the alternative boarding point and the geographic position information of the client.

a 2: whether the alternative boarding point and the service request end are positioned on the same side of the same road section or not.

a 3: and the road grade of the road section where the alternative boarding point is located.

Here, in the urban road class, there are included: the expressway, the main road, the secondary main road, the branch road and the street lane are in five grades. Generally, the road grade can represent the difficulty degree of reaching between the alternative boarding point and the service request end to a certain extent; for example, if the road grade of the road segment where the candidate pick-up point is located is an express road, a main road, and a sub-main road, it is difficult for the user to reach the other side from one side of the road, and therefore if the candidate pick-up point and the service request end are located on different sides of the same road segment and the road grade of the road segment is too high, it is obviously unreasonable to use the candidate pick-up point as a pick-up point shown to the user, and therefore, the candidate pick-up point is used as one of the correlation characteristics as an input of the correlation detection model.

a 4: the alternative pick-up point is used as the first heat value of the selected pick-up point.

Here, the selection of the pick-up point means that the pick-up point for indicating a specific pick-up position of the service provider is selected when the user issues an order based on each pick-up point displayed to the user on the target service page.

The first heat value may be obtained in the following manner:

acquiring order information of a historical order; the order information comprises selected boarding point information; and counting a first quantity of historical orders with the alternative boarding points as the selected boarding points based on the selected boarding point information, and taking the first quantity as a first heat value.

Here, the order information of the historical order includes the selected boarding point information, and also includes other information, such as an order ID, an order issuing time, a service request terminal identifier, a service requester account, and the like. The server can determine a first quantity of all history orders taking the alternative boarding points as the selected boarding points based on the selected boarding point information corresponding to each history order, and the first quantity is used as a first heat value.

The greater the first number of historical orders for which an alternative pick-up point is selected as the selected pick-up point, the greater the heat characterizing the alternative pick-up point.

Here, unlike the third popularity, as long as the alternative pick-up point appears as the selected pick-up point in the history order, the history order is counted as the counted object regardless of whether the information of the alternative pick-up point is part of the information of the displayed pick-up point.

a 5: a second heat value of the road segment on which the alternative boarding point is located.

Here, the second calorific value may be acquired in the following manner:

obtaining order information of a historical order and determining a vehicle getting-on point on the same road section of the same road section with the alternative vehicle getting-on point; the order information comprises selected boarding point information;

according to the fact that the order information comprises the selected getting-on point information, a first quantity of historical orders with the alternative getting-on points as the selected getting-on points is counted, and a second quantity of historical orders with the getting-on points on the same road section as the selected getting-on points is counted; the sum of the first quantity and the second quantity is taken as a second calorific value.

a 6: a third heating value selected after the alternative pick-up point is presented to the user.

Here, the third calorific value may be acquired in the following manner:

acquiring order information of a historical order; the order information includes: displaying the getting-on point information and selecting a getting-on point message; according to the displayed boarding point information and the selected boarding point information, counting a third quantity of historical orders when the alternative boarding points are used as the selected boarding points; the third amount is taken as a third calorific value.

Here, the display boarding point is also called a fixed point boarding point, and refers to a boarding point which is determined for the user and displayed to the user after sequencing all the alternative boarding points. The displaying of the boarding point information comprises the following steps: at least one display pick-up point geographical location coordinates and pick-up point name. In contrast to the second heat value, when the third quantity is determined, the historical order is counted as a counted object only when the display pick-up point information includes alternative pick-up point information and the alternative pick-up point is used as a selected pick-up point in the historical order.

After the target characteristic values of the candidate boarding points under the multiple correlation characteristics are determined, the target characteristic values of the candidate boarding points can be input into a correlation detection model, and the correlation between the candidate boarding points and the service request end is obtained.

The correlation detection model is formed by training a plurality of first sample characteristic values of the vehicle-mounted points under a plurality of correlation characteristics and a plurality of second sample characteristic values under a plurality of noise characteristics.

Referring to fig. 3, an embodiment of the present application further provides a specific method for training a correlation detection model, including:

s301: a plurality of sample pick-up points are determined.

In a specific implementation, the sample loading points include a positive sample loading point and a negative sample loading point.

In the historical order, the boarding point displayed to the user by the service request terminal is called a displayed boarding point, and the boarding point selected as the boarding point by the user when the user issues the order is called a selected boarding point. The display of the boarding point refers to the boarding point displayed to the user in the alternative boarding points which are firstly determined by the server for the user according to the geographical position information of the service request end; the selected pick-up point may be one of these show pick-up points, or other pick-up points that the user has determined are not included in the show pick-up points, either by a manual precision search, or by the user dragging a map shown on the service page. In this embodiment of the application, the selected getting-on point may be used as a positive sample getting-on point, and the other getting-on points except the selected getting-on point in the showing getting-on points may be used as negative sample getting-on points, so that the numbers of the positive sample getting-on points and the negative sample getting-on points are kept in a certain proportion.

The real correlation degree between the positive sample getting-on point and the service request end in the corresponding historical order is 1, and the real correlation degree between the negative sample getting-on point and the service request end in the corresponding historical order is 0.

Here, the plurality of sample boarding points may be determined in the following manner: obtaining sample order information of a plurality of sample orders; the sample order information includes: a sample display boarding point and a sample selection boarding point;

determining a selected getting-on point of the sample as a positive sample getting-on point;

and determining the other vehicle-entering points except the selected vehicle-entering point of the sample in the vehicle-entering points of the sample display as the vehicle-entering points of the negative sample.

S302: and acquiring first sample characteristic values of the vehicle-mounted points on the samples under a plurality of correlation characteristics and second sample characteristic values under a plurality of noise characteristics.

Here, the plurality of correlation features corresponding to the cart-on point on the sample are the same as the plurality of correlation features corresponding to the alternative cart-on point, including but not limited to one or more of the following:

the distance between the sample getting-on point and the service request end, whether the sample getting-on point and the service request end are located on the same side of the same road section, the road grade of the road section where the sample getting-on point is located, a first heat value of the sample getting-on point as the selected getting-on point, a second heat value of the road section where the sample getting-on point is located, and a third heat value of the sample getting-on point selected after being displayed to the user.

It should be noted here that there are multiple historical orders corresponding to each sample pick-up point. When the correlation characteristics comprise at least one of the distance between the sample getting-on point and the service request end, whether the sample getting-on point and the service request end are positioned on the same side of the same road section, and the road grade of the road section on which the sample getting-on point is positioned, for a plurality of historical orders corresponding to the sample getting-on point, the correlation information corresponding to each historical order can form a sample data corresponding to the sample getting-on point. That is, a vehicle point on one sample corresponds to a plurality of sample data; these sample data are independent of each other. One sample data is used as an input in the model training process.

The noise characteristics refer to characteristics that adversely affect the accuracy of the correlation detection model, such as the degree of familiarity of the user with the location where the user is located, the degree of familiarity of the driver with the location where the driver wants to take over driving, and the like. These are generally posterior features, i.e. features whose characteristic values can only be determined after the vehicle-entering points have been sorted. Here, other objective values are used to reflect the familiarity laterally. Further, the noise characteristics include, but are not limited to, one or more of the following b 1-b 3:

b 1: and the sample boarding point is used as the instant call duration between the service request end and the service providing end when the selected boarding point is selected.

b 2: and (5) connecting the driving route.

b 3: and (5) receiving driving time.

In addition, the historical frequency of the service request terminal appearing near the position corresponding to the current geographical position information can be determined based on the geographical position information of the departure place in the historical order corresponding to the service request terminal and the current geographical position information of the service request terminal, and the historical frequency is also used as a part of the noise characteristic. The more this history, the more familiar the service requester is to the environment.

S303: constructing a plurality of correlation decision trees corresponding to the correlation characteristics based on the first sample characteristic values, and constructing a plurality of noise decision trees corresponding to the noise characteristics based on the second sample characteristic values; and taking the constructed multiple relevance decision trees as a relevance detection model.

Wherein, the correlation decision tree and the noise decision tree are alternately constructed; the residual error of any one of the correlation decision trees is the fitting target of the next noise decision tree with the closest distance to the correlation decision tree, and the residual error of any one of the noise decision trees is the fitting target of the next correlation decision tree with the closest distance to the noise decision tree.

Here, the alternating construction of the correlation decision tree and the noise decision tree means that after one correlation decision tree is constructed, one noise decision tree is constructed by using a residual error corresponding to the correlation decision tree as a fitting target, then a next correlation decision tree … … is constructed by using the residual error of the newly constructed noise decision tree as the fitting target, so that a plurality of correlation decision trees are obtained by the alternating construction of the correlation decision trees and the noise decision trees, and the constructed plurality of correlation decision trees are used as the correlation detection model.

Inputting a first sample characteristic value of a sample getting-on point under the correlation characteristic and a second sample characteristic value under a plurality of noise characteristics into a constructed correlation decision tree to obtain a first prediction correlation degree corresponding to each sample getting-on point; calculating the difference between a first prediction correlation degree corresponding to each sample vehicle-loading point and a real correlation degree for each sample vehicle-loading point; the first difference is a first residual error corresponding to the vehicle point on the sample; and the first residual errors corresponding to the vehicle-mounted points on each sample are the residual errors corresponding to the relevance decision tree.

Similarly, inputting a first sample characteristic value of the sample getting-on point under the correlation characteristic and a second sample characteristic value under a plurality of noise characteristics into the constructed noise decision tree to obtain a second prediction correlation corresponding to each sample getting-on point; calculating a second difference value between a second prediction correlation degree corresponding to the vehicle-loading point on the sample and the real correlation degree aiming at the vehicle-loading point on the sample; the second difference is a second residual error corresponding to the vehicle point on the sample; and the second residual errors corresponding to the vehicle-mounted points on each sample are the residual errors corresponding to the noise decision tree.

The following describes a process of alternately constructing a correlation decision tree and a noise decision tree by taking the first constructed decision tree as the correlation decision tree.

Specifically, the specific process for alternately constructing the correlation decision tree and the noise decision tree provided by the embodiment of the present application is as follows:

a: constructing a first relevance decision tree:

step 1: a current target relevance feature is selected from the relevance features as a parent node of the relevance decision tree.

Here, first, an information gain and an information gain ratio corresponding to each correlation characteristic are calculated from the first sample characteristic value of the vehicle point on each sample under each correlation characteristic.

Wherein the information gain of each correlation feature satisfies the formula:

g(D,A)＝H(D)-H(D|A)。

wherein A is a correlation characteristic, and D is a set formed by vehicle-loading points of each sample; g (D, A) is the information gain of the correlation characteristic A to the set D; h (D) is the empirical entropy of D, and H (D | A) is the empirical conditional entropy of the set D given the correlation characteristic A.

The information of each correlation characteristic once satisfies the formula:

wherein A is a correlation characteristic, and D is a set formed by vehicle-loading points of each sample; g_R(D, A) is the information gain ratio of feature A to set D, and g (D, A) is the information gain of correlation feature A to set D; h_A(D) Is the entropy of the set D with respect to the values of the relevance feature a.

Then eliminating the correlation characteristics of which the information gain is lower than the average value of the information gains of all the correlation characteristics;

and finally, arranging the rest correlation characteristics according to the sequence of the information gain ratio from large to small, and taking the correlation characteristic with the maximum information gain ratio as the selected current target correlation characteristic.

Step 2: and dividing each sample getting-on point into each characteristic value interval corresponding to the parent node according to the plurality of characteristic intervals corresponding to the parent node and the first sample characteristic value of each sample getting-on point under the current target correlation characteristic.

Here, the characteristic value interval may be specifically divided according to actual needs, and the dividing manner is not described herein again.

And step 3: detecting whether the depth of the correlation decision tree meets a preset depth or not; if not, executing the step 4; if so, step 6 is performed.

And 4, step 4: for each interval of features, a new current target relevance feature is determined from the relevance features other than the relevance feature that has been selected as the current target relevance feature, as a child node of the parent node.

Here, the selection manner of the new current target correlation feature is similar to that in step 1, and is not described herein again.

And 5: taking the child node as a new parent node; and returning to the step 2;

step 6: and obtaining the trained first relevance decision tree.

And 7: and calculating a first prediction correlation corresponding to the vehicle points on each sample according to the first correlation decision tree.

And 8: and obtaining first residual errors respectively corresponding to the sample upper vehicle points according to the real correlation corresponding to the sample upper vehicle points contained in each leaf node in the first correlation decision tree and the prediction correlation respectively corresponding to the sample upper vehicle points.

Illustratively, there are, for example, 5 sample loading points A1-A5, where A1-A3 are positive sample loading points and A4 and A5 are negative sample loading points. The true correlation degrees corresponding to a 1-A3 are 1, and the true correlation degrees corresponding to a 4-a 5 are 0.

After the feature values of the 5 sample boarding points A1-A5 under the target correlation characteristics corresponding to each first correlation decision tree are input into the first correlation decision tree, the prediction correlation degrees corresponding to the sample boarding points are respectively: 0.7, 0.6, 0.8, 0.2, and 0.3, the first residuals corresponding to a 1-a 5, respectively, are: 0.3, 0.4, 0.2, -0.8 and-0.7.

B: constructing a first noise decision tree according to the first residual error of the vehicle-mounted point on each sample corresponding to the first correlation decision tree:

and step 9: a current target noise feature is selected from the noise features as a parent node of the first noise decision tree.

Step 10: and dividing each sample getting-on point into each characteristic value interval corresponding to the parent node according to the plurality of characteristic intervals corresponding to the parent node and the second sample characteristic value of each sample getting-on point under the current target noise characteristic.

Step 11: detecting whether the depth of a first noise decision tree meets a preset depth; if not, executing step 12; if so, step 14 is performed.

Step 12: for each feature interval, a new current target noise feature is determined from the noise features other than the noise feature that has been selected as the current target noise feature as a child node of the parent node.

Step 13: taking the child node as a new parent node; and returning to the step 10;

step 14: and obtaining a first trained noise decision tree.

Step 15: and calculating second prediction correlation corresponding to the vehicle points on each sample according to the first noise decision tree.

Step 16: and obtaining second residual errors respectively corresponding to the vehicle-mounted points on the samples according to the first residual errors respectively corresponding to the vehicle-mounted points on the samples obtained in the step 8 and the second prediction correlation corresponding to the vehicle-mounted points on the samples.

And step 17: and obtaining the prediction correlation degree of the vehicle-mounted point on each sample according to the first prediction correlation degree and the second prediction correlation degree corresponding to the vehicle-mounted point on each sample.

Step 18: and calculating loss according to the predicted correlation degree and the real correlation degree of the vehicle points on each sample. And if the loss is greater than a preset loss threshold value, constructing a second relevance decision tree based on a second residual error corresponding to the vehicle point on each sample.

C: and constructing a second correlation decision tree according to the second residual error of the vehicle-mounted point on each sample corresponding to the first noise decision tree:

step 19: a current target relevance feature is selected from the relevance features as a parent node of the second relevance decision tree.

Step 20: and dividing each sample getting-on point into each characteristic value interval corresponding to the parent node according to the plurality of characteristic intervals corresponding to the parent node and the first sample characteristic value of each sample getting-on point under the current target correlation characteristic.

Step 21: detecting whether the depth of the second correlation decision tree meets a preset depth; if not, go to step 22; if so, step 24 is performed.

Step 22: for each interval of features, a new current target relevance feature is determined from the relevance features other than the relevance feature that has been selected as the current target relevance feature, as a child node of the parent node.

Step 23: taking the child node as a new parent node; and returning to step 20;

step 24: and obtaining a second trained relevance decision tree.

Step 25: and calculating a third prediction correlation corresponding to the vehicle points on each sample according to the second correlation decision tree.

Step 26: and obtaining third residual errors respectively corresponding to the vehicle-mounted points on the samples according to the second residual errors respectively corresponding to the vehicle-mounted points on the samples obtained in the step 16 and the third prediction correlation corresponding to the vehicle-mounted points on the samples.

Step 27: and obtaining the prediction correlation degree of the vehicle-mounted point on each sample according to the first prediction correlation degree, the second prediction correlation degree and the third prediction correlation degree corresponding to the vehicle-mounted point on each sample.

Step 28: and calculating loss according to the predicted correlation and the real correlation of the vehicle points on each sample. And if the loss is greater than a preset loss threshold value, constructing a second noise decision tree based on a third residual error corresponding to the vehicle point on each sample.

……

And circulating until the loss is not greater than a preset loss threshold value.

Finally, a plurality of correlation decision trees and a plurality of noise decision trees are formed.

Then, the formed multiple relevance decision trees are used as the relevance detection model in the embodiment of the present application.

And inputting the target characteristic values of the alternative nodes under the multiple correlation characteristics into a pre-trained correlation degree detection model, namely inputting the target characteristic values of the alternative nodes under the target correlation characteristics corresponding to the correlation decision tree into the corresponding correlation decision tree. Each relevance decision tree can determine a sub-relevance for the alternative boarding points; and then obtaining the correlation between the alternative vehicle-entering points and the service request end based on the sub-correlation determined by all the correlation decision trees for the alternative vehicle-entering points.

III: in step S203, when the candidate pick-up points are sorted according to the correlation degrees corresponding to the candidate pick-up points, the higher the correlation degree is, the further the sorting is.

In addition, in another embodiment of the present application, after receiving the step S203 and sorting the pick-up points, the method further includes:

s204: and determining a preset number of fixed point boarding points from the alternative boarding points according to the sequence of the alternative boarding points, and displaying the fixed point boarding points through the service request end.

Here, when the fixed point boarding point is displayed through the service request end, if the execution main body of the method is the service request end, the service request end directly displays the fixed point boarding point on the service interface; if the execution main body of the method is the server, the server can send the fixed point boarding point to the service request end; and when the service request end receives the fixed point boarding point sent by the server, the fixed point boarding point is displayed on the corresponding service interface.

According to the method and the device, a plurality of candidate getting-on points are obtained through geographic position information of a service request end, target characteristic values of the candidate getting-on points under a plurality of correlation characteristics are input into a pre-trained correlation degree detection model aiming at each candidate getting-on point, correlation degrees between each candidate getting-on point and the service request end are obtained, and the candidate getting-on points are sorted based on the correlation degrees corresponding to the candidate getting-on points; the relevancy detection model comprises a plurality of relevancy decision trees; the method comprises the steps that a plurality of correlation decision trees and a plurality of noise decision trees are alternately trained, in the training process, the residual error of any one correlation decision tree is the fitting target of the next noise decision tree closest to the any one correlation decision tree, the residual error of any one noise decision tree is the fitting target of the next correlation decision tree closest to the any one noise decision tree, in the using process, only the correlation decision tree is used as a correlation detection model to obtain the correlation of each alternative boarding point, in the training process of the correlation detection model, the influence of correlation characteristics and noise characteristics on the model is considered, in the using process, the influence of the noise characteristics on correlation detection results is eliminated, the higher correlation detection precision is achieved, and in the sorting process of the boarding points, the higher sorting accuracy is achieved.

Based on the same inventive concept, the embodiment of the present application further provides a device for sorting vehicle-entering points corresponding to the method for sorting vehicle-entering points, and since the principle of solving the problem of the device in the embodiment of the present application is similar to that of the method for sorting vehicle-entering points in the embodiment of the present application, the implementation of the device can refer to the implementation of the method, and repeated details are omitted.

Referring to fig. 4, a schematic diagram of an apparatus for sorting boarding points according to an embodiment of the present application is shown, where the apparatus includes: a first obtaining module 41, a first determining module 42, a first ordering module 43; wherein the content of the first and second substances,

a first obtaining module 41, configured to obtain geographic location information of a service request end, and determine a plurality of candidate boarding points based on the geographic location information;

a first determining module 42, configured to, for each candidate boarding point, input a target feature value of the candidate boarding point under multiple correlation features into a pre-trained correlation detection model, and obtain a correlation between the candidate boarding point and a service request end;

the first sequencing module 43 is configured to sequence the alternative boarding points according to the respective corresponding relevancy of each alternative boarding point, so that a user selects a final boarding point according to a sequencing result;

the relevancy detection model comprises a plurality of relevancy decision trees; and training a plurality of correlation decision trees and a plurality of noise decision trees alternately, wherein during training, the residual error of any one correlation decision tree is the fitting target of the next noise decision tree with the closest distance to the any correlation decision tree, and the residual error of any one noise decision tree is the fitting target of the next correlation decision tree with the closest distance to the any noise decision tree.

In a possible embodiment, the method further comprises: a first training module 44, configured to train the correlation detection model in the following manner:

determining a plurality of sample loading points;

acquiring first sample characteristic values of vehicle-entering points of each sample under a plurality of correlation characteristics and second sample characteristic values under a plurality of noise characteristics;

constructing a plurality of correlation decision trees corresponding to the correlation characteristics based on the first sample characteristic values, and constructing a plurality of noise decision trees corresponding to the noise characteristics based on the second sample characteristic values;

and taking the constructed multiple relevance decision trees as a relevance detection model.

In one possible embodiment, the sample loading station comprises: a positive sample loading point and a negative sample loading point;

the first training module 44 is specifically configured to determine a plurality of sample boarding points by:

In a possible embodiment, the first obtaining module 41 is configured to obtain geographic location information of the service request end and determine a plurality of candidate boarding points based on the geographic location information, by:

receiving geographic position information sent by a service request end when a target service interface is opened;

based on the geographic position information, obtaining a plurality of alternative boarding points corresponding to the geographic position information from a boarding point database;

and the distance between the geographical position indicated by the geographical position information and the alternative boarding point is smaller than a preset distance threshold value.

In one possible embodiment, the correlation features include one or more of: the distance between the alternative getting-on point and the service request end, whether the alternative getting-on point and the service request end are located on the same side of the same road section, the road grade of the road section where the alternative getting-on point is located, a first heat value of the selected getting-on point where the alternative getting-on point is located, a second heat value of the road section where the alternative getting-on point is located, and a third heat value of the alternative getting-on point which is selected after being displayed to the user.

In a possible implementation, the first obtaining module 41 is further configured to:

aiming at the condition that the correlation characteristics comprise the distance between the alternative boarding point and the service request end, the distance between the alternative boarding point and the service request end is obtained by adopting the following mode:

obtaining the distance between the alternative boarding point and the service request end according to the geographic position coordinate corresponding to the alternative boarding point and the geographic position information of the service request end;

for the case that the correlation characteristic includes the first heat value, the first heat value is obtained in the following manner:

acquiring order information of a historical order; the order information comprises selected boarding point information; counting a first quantity of historical orders taking alternative boarding points as the selected boarding points based on the selected boarding point information, and taking the first quantity as a first heat value;

for the case that the correlation characteristic includes the second heat value, the second heat value is obtained in the following manner:

according to the fact that the order information comprises the selected getting-on point information, a first quantity of historical orders with the alternative getting-on points as the selected getting-on points is counted, and a second quantity of historical orders with the getting-on points on the same road section as the selected getting-on points is counted; taking the sum of the first quantity and the second quantity as a second heat value;

for the case that the correlation characteristic includes a third heat value, the third heat value is obtained in the following manner:

acquiring order information of a historical order; the order information includes: displaying the information of the boarding points and selecting the information of the boarding points;

according to the displayed boarding point information and the selected boarding point information, counting a third quantity of historical orders when the alternative boarding points are used as the selected boarding points; the third amount is taken as a third calorific value.

In one possible embodiment, the noise signature includes one or more of: and taking the sample boarding point as the instant call duration, the driving receiving distance and the driving receiving time between the service request end and the service providing end when the selected boarding point is selected.

In a possible embodiment, the method further comprises: and a first display module 45, configured to determine a preset number of fixed-point boarding points from the alternative boarding points according to the sequence of the alternative boarding points, and display the fixed-point boarding points through the service request end.

The description of the processing flow of each module in the device and the interaction flow between the modules may refer to the related description in the above method embodiments, and will not be described in detail here.

An embodiment of the present application further provides a computer device 50, as shown in fig. 5, which is a schematic structural diagram of the computer device 50 provided in the embodiment of the present application, and includes: a processor 51, a memory 52, and a bus 53. The memory 52 stores machine-readable instructions executable by the processor 51 (for example, the execution instructions corresponding to the first obtaining module 41, the first determining module 42, the first ordering module 43 in the apparatus in fig. 4, etc.), when the computer device 50 is running, the processor 51 communicates with the memory 52 via the bus 53, and the machine-readable instructions, when executed by the processor 51, perform the following processes:

acquiring geographic position information of a service request end, and determining a plurality of alternative boarding points based on the geographic position information;

for each alternative boarding point, inputting a target characteristic value of the alternative boarding point under a plurality of correlation characteristics into a pre-trained correlation detection model, and acquiring the correlation between the alternative boarding point and a service request end;

sequencing the alternative boarding points according to the corresponding correlation degree of each alternative boarding point so that a user can select a final boarding point according to a sequencing result;

In one possible embodiment, the processor 51 executes instructions that train the correlation detection model in the following manner:

determining a plurality of sample loading points;

In one possible embodiment, the instructions executed by the processor 51 include, for a sample pick-up point: a positive sample loading point and a negative sample loading point;

determining a plurality of sample boarding points, specifically comprising:

In a possible embodiment, the processor 51 executes instructions to obtain geographic location information of the service request end, and determine a plurality of candidate boarding points based on the geographic location information, including:

In one possible embodiment, the processor 51 executes instructions in which the dependency characteristics include one or more of the following: the distance between the alternative getting-on point and the service request end, whether the alternative getting-on point and the service request end are located on the same side of the same road section, the road grade of the road section where the alternative getting-on point is located, a first heat value of the selected getting-on point where the alternative getting-on point is located, a second heat value of the road section where the alternative getting-on point is located, and a third heat value of the alternative getting-on point which is selected after being displayed to the user.

In one possible embodiment, the processor 51 executes instructions to obtain the distance between the candidate boarding point and the service request end in the following manner for the case that the correlation characteristic includes the distance between the candidate boarding point and the service request end:

In one possible embodiment, the processor 51 executes instructions in which the noise characteristics include one or more of: and taking the sample boarding point as the instant call duration, the driving receiving distance and the driving receiving time between the service request end and the service providing end when the selected boarding point is selected.

In a possible embodiment, after the instruction executed by the processor 51 sorts the candidate pick-up points according to the degree of correlation corresponding to each candidate pick-up point, the method further includes:

and determining a preset number of fixed point boarding points from the alternative boarding points according to the sequence of the alternative boarding points, and displaying the fixed point boarding points through the service request end.

The embodiment of the present application further provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps of the above-mentioned boarding point ordering method are performed.

Specifically, the storage medium can be a general storage medium, such as a mobile disk, a hard disk, and the like, and when a computer program on the storage medium is executed, the above method for ordering the boarding points can be executed, so that the problem of low ordering accuracy when the boarding points are ordered in the prior art is solved, and the effect of improving the ordering accuracy of the boarding points is achieved.

Based on the same inventive concept, referring to fig. 6, an embodiment of the present application further provides an information ordering method, including:

s601: acquiring user information, and determining a plurality of pieces of alternative information corresponding to the user information based on the user information;

s602: for each piece of alternative information, inputting a target characteristic value of the alternative information under a plurality of correlation characteristics into a pre-trained correlation detection model, and acquiring the correlation between the alternative information and user information;

s603: sorting the alternative information according to the corresponding relevancy of each alternative information, so that a user can select the final alternative information according to a sorting result;

Wherein, the correlation detection model is trained by adopting the following method:

determining a plurality of pieces of sample information;

acquiring a first sample characteristic value of each piece of sample information under a plurality of correlation characteristics and a second sample characteristic value under a plurality of noise characteristics;

According to the method and the device, the target characteristic values of a plurality of pieces of alternative information corresponding to the user information under a plurality of correlation characteristics are input into a pre-trained correlation degree detection model, the correlation degree between each piece of alternative information and the user information is obtained, and then the alternative information is ranked based on the correlation degree corresponding to each piece of alternative information; the relevancy detection model comprises a plurality of relevancy decision trees; the method comprises the steps that a plurality of correlation decision trees and a plurality of noise decision trees are alternately trained, in the training process, the residual error of any one of the correlation decision trees is the fitting target of the next noise decision tree closest to the any one of the correlation decision trees, the residual error of any one of the noise decision trees is the fitting target of the next correlation decision tree closest to the any one of the noise decision trees, in the using process, only the correlation decision tree is used as a correlation detection model to obtain the correlation degree of each piece of alternative information, further, in the process of training the correlation detection model, the influence of correlation characteristics and noise characteristics on the model is considered, in the using process, the influence of the noise characteristics on correlation detection results is eliminated, the correlation detection precision is higher, and further, when the alternatives are ranked, the ranking accuracy is higher.

Referring to fig. 7, an embodiment of the present application further provides an information sorting apparatus, including:

a second obtaining module 71, configured to obtain the user information, and determine, based on the user information, a plurality of pieces of alternative information corresponding to the user information;

a second determining module 72, configured to, for each piece of candidate information, input a target feature value of the candidate information under multiple correlation features into a pre-trained correlation detection model, and obtain a correlation between the candidate information and user information;

the second sorting module 73 is configured to sort the alternative information according to the degree of correlation corresponding to each piece of alternative information, so that a user selects final alternative information according to a sorting result;

In an alternative embodiment, the method further comprises: a second training module 74 for training the correlation detection model in the following manner:

determining a plurality of pieces of sample information;

An embodiment of the present application further provides a computer device 80, as shown in fig. 8, which is a schematic structural diagram of the computer device 80 provided in the embodiment of the present application, and includes: a processor 81, a memory 82, and a bus 83. The memory 82 stores machine-readable instructions executable by the processor 81 (for example, the execution instructions corresponding to the second obtaining module 71, the second determining module 72, the second sorting module 73, etc. in the apparatus in fig. 7), when the computer device 80 is running, the processor 81 and the memory 82 communicate via the bus 83, and when the processor 81 executes the following processing:

acquiring user information, and determining a plurality of pieces of alternative information corresponding to the user information based on the user information;

for each piece of alternative information, inputting a target characteristic value of the alternative information under a plurality of correlation characteristics into a pre-trained correlation detection model, and acquiring the correlation between the alternative information and user information;

sorting the alternative information according to the corresponding relevancy of each alternative information, so that a user can select the final alternative information according to a sorting result;

In one possible embodiment, the processor 81 executes instructions that train the correlation detection model in the following manner:

determining a plurality of pieces of sample information;

The embodiment of the application also provides a computer readable storage medium, wherein a computer program is stored on the computer readable storage medium, and when the computer program is executed by a processor, the steps of the method for ordering the information are executed.

Specifically, the storage medium can be a general storage medium, such as a mobile disk, a hard disk, and the like, and when a computer program on the storage medium is executed, the method for ordering information can be executed, so that the problem of low information ordering accuracy in the prior art is solved, and the effect of improving the information ordering accuracy is achieved.

It can be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working processes of the system and the apparatus described above may refer to corresponding processes in the method embodiments, and are not described in detail in this application. In the several embodiments provided in the present application, it should be understood that the disclosed system, apparatus and method may be implemented in other ways. The above-described apparatus embodiments are merely illustrative, and for example, the division of the modules is merely a logical division, and there may be other divisions in actual implementation, and for example, a plurality of modules or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection of devices or modules through some communication interfaces, and may be in an electrical, mechanical or other form.

The modules described as separate parts may or may not be physically separate, and parts displayed as modules may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit.

The functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a non-volatile computer-readable storage medium executable by a processor. Based on such understanding, the technical solution of the present application or portions thereof that substantially contribute to the prior art may be embodied in the form of a software product stored in a storage medium and including instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present application. And the aforementioned storage medium includes: various media capable of storing program codes, such as a U disk, a removable hard disk, a ROM, a RAM, a magnetic disk, or an optical disk.

The above description is only for the specific embodiments of the present application, but the scope of the present application is not limited thereto, and any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present application, and shall be covered by the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A method of pick-up ordering, comprising:

for each alternative boarding point, inputting a target characteristic value of the alternative boarding point under a plurality of correlation characteristics into a pre-trained correlation detection model, and acquiring the correlation between the alternative boarding point and the service request end;

2. The method of claim 1, wherein the correlation detection model is trained by:

determining a plurality of sample loading points;

obtaining a first sample characteristic value of each sample getting-on point under a plurality of correlation characteristics and a second sample characteristic value under a plurality of noise characteristics;

constructing a plurality of correlation decision trees corresponding to the correlation features based on the first sample feature values, and constructing a plurality of noise decision trees corresponding to the noise features based on the second sample feature values;

and taking the constructed multiple relevance decision trees as the relevance detection model.

3. The method of claim 2, wherein the sample pick-up comprises: a positive sample loading point and a negative sample loading point;

the determining the plurality of sample getting-on points specifically includes:

4. The method of claim 1, wherein the obtaining geographic location information of the service request end and determining a plurality of alternative pick-up points based on the geographic location information comprises:

receiving the geographical position information sent by a service request end when a target service interface is opened;

and the distance between the geographical position indicated by the geographical position information and the alternative vehicle-entering point is smaller than a preset distance threshold value.

5. The method of claim 1, wherein the correlation features comprise one or more of: the distance between the alternative vehicle getting-on point and the service request end, whether the alternative vehicle getting-on point and the service request end are located on the same side of the same road section, the road grade of the road section where the alternative vehicle getting-on point is located, the first heat value of the road section where the alternative vehicle getting-on point is used as the selected vehicle getting-on point, the second heat value of the road section where the alternative vehicle getting-on point is located, and the third heat value of the alternative vehicle getting-on point selected after being displayed to the user.

6. The method according to claim 5, wherein for the case that the correlation characteristic comprises the distance between the candidate boarding point and the service request end, the distance between the candidate boarding point and the service request end is obtained by adopting the following way:

for the case that the correlation feature includes the first heat value, the first heat value is obtained in the following manner:

acquiring order information of a historical order; the order information comprises selected boarding point information; counting a first quantity of historical orders taking the alternative boarding points as the selected boarding points based on the selected boarding point information, and taking the first quantity as the first heat value;

for the case that the correlation feature includes the second heat value, the second heat value is obtained in the following manner:

obtaining order information of a historical order and determining the vehicle-entering points on the same road section of the same road section with the alternative vehicle-entering points; the order information comprises selected boarding point information;

according to the fact that the order information comprises selected boarding points information, counting a first quantity of historical orders taking the alternative boarding points as the selected boarding points, and counting a second quantity of the historical orders taking the boarding points on the same road section as the selected boarding points; taking the sum of the first amount and the second amount as the second heat value;

for a case that the correlation feature includes the third heat value, obtaining the third heat value in the following manner:

acquiring order information of a historical order; the order information comprises: displaying the information of the boarding points and selecting the information of the boarding points;

according to the displayed pick-up point information and the selected pick-up point information, counting a third quantity of historical orders when the alternative pick-up points are used as the selected pick-up points; taking the third amount as the third heat value.

7. The method of claim 2, wherein the noise signature comprises one or more of: and when the sample boarding point is taken as the selected boarding point, the instant call duration, the driving receiving distance and the driving receiving time between the service request end and the service providing end are determined.

8. The method according to claim 1, wherein after sorting the candidate pick-up points according to the degree of correlation corresponding to each candidate pick-up point, the method further comprises:

9. A method of ordering information, comprising:

for each piece of alternative information, inputting the target characteristic value of the alternative information under a plurality of correlation characteristics into a pre-trained correlation detection model, and acquiring the correlation between the alternative information and the user information;

sorting the alternative information according to the corresponding relevancy of each alternative information, so that a user can select final alternative information according to a sorting result;

10. An apparatus for boarding point sequencing, comprising:

the first sequencing module is used for sequencing the alternative vehicle-entering points according to the corresponding relevancy of each alternative vehicle-entering point; so that the user can select the final boarding point according to the sorting result;

11. An apparatus for sorting information, comprising:

12. A computer device, comprising: a processor, a storage medium and a bus, the storage medium storing machine-readable instructions executable by the processor, the processor and the storage medium communicating via the bus when the computer device is running, the processor executing the machine-readable instructions to perform the steps of the method according to any one of claims 1 to 9.

13. A computer-readable storage medium, having stored thereon a computer program which, when being executed by a processor, is adapted to carry out the steps of the method according to any one of claims 1 to 9.