CN111461380A

CN111461380A - Position prediction method and device

Info

Publication number: CN111461380A
Application number: CN201910055395.8A
Authority: CN
Inventors: 谢君; 卓呈祥
Original assignee: Beijing Didi Infinity Technology and Development Co Ltd
Current assignee: Beijing Didi Infinity Technology and Development Co Ltd
Priority date: 2019-01-21
Filing date: 2019-01-21
Publication date: 2020-07-28
Also published as: WO2020151725A1

Abstract

The embodiment of the application provides a position prediction method and a position prediction device, wherein the method comprises the following steps: acquiring a first characteristic value of a user to be predicted under at least one historical behavior characteristic, and a second characteristic value and a third characteristic value of the user to be predicted under a target time characteristic and a target position characteristic in a future preset time period; and inputting the first characteristic value, the second characteristic value and the third characteristic value into a position prediction model trained in advance, and obtaining a position prediction result of the user to be predicted in the future preset time period. The method and the device for predicting the user position can determine the position of the user to be predicted in the future preset time period with higher accuracy, so that the user position information can be predicted in advance, and resource allocation and configuration of relevant service strategies can be facilitated in advance.

Description

Position prediction method and device

Technical Field

The present application relates to the field of machine learning technologies, and in particular, to a position prediction method and apparatus.

Background

By acquiring the user position information, personalized services, incentive strategies, capacity allocation strategies and the like related to the user position can be matched for the user. For example, by predicting the city where the user is located the next day, the user who travels in different places or goes on business can be found, differential operation is performed, and user experience is improved.

Currently, when the user position is determined, the user positioning position information can be obtained in real time, but the timeliness of the method for obtaining the positioning position in real time is poor, and resource allocation and configuration of related service strategies are not facilitated in advance.

Disclosure of Invention

In view of the above, an object of the embodiments of the present application is to provide a method and an apparatus for location prediction, which can determine, with higher accuracy, a location of a user to be predicted within a future preset time period.

In a first aspect, an embodiment of the present application provides a location prediction method, where the method includes:

acquiring a first characteristic value of a user to be predicted under at least one historical behavior characteristic, and a second characteristic value and a third characteristic value of the user to be predicted under a target time characteristic and a target position characteristic in a future preset time period;

and inputting the first characteristic value, the second characteristic value and the third characteristic value into a position prediction model trained in advance, and obtaining a position prediction result of the user to be predicted in the future preset time period.

In an optional embodiment, the obtaining a first feature value of the user to be predicted under at least one historical behavior feature includes:

extracting the area historical behavior information of the user to be predicted in the area to be predicted from at least one type of historical behavior information of the user to be predicted, and extracting the area time historical behavior information of the user to be predicted in each historical time period in a plurality of historical time periods;

determining a first characteristic value under a first historical behavior characteristic of the user to be predicted according to the at least one type of historical behavior information of the user to be predicted, the regional historical behavior information and the regional time historical behavior information;

and determining a first characteristic value under a second historical behavior characteristic of the user to be predicted according to the at least one type of historical behavior information of the user to be predicted.

In an alternative embodiment, the first historical behavior feature comprises one or more of:

the occurrence frequency of the historical behaviors corresponding to the POI classifications, the occurrence frequency of the historical behaviors in a preset time period, the occurrence frequency of the historical behaviors in a working day, the occurrence frequency of the historical behaviors in a non-working day, and the occurrence frequency corresponding to different historical behaviors.

In an alternative embodiment, the second historical behavior characteristic comprises one or more of:

whether the area where the last historical behavior occurs is an area to be predicted or not, the time interval between the time when the last historical behavior occurs and the time to be predicted, the POI classification of the destination of the last historical behavior, the POI classification of the departure place of the last historical behavior, and the number of areas which the user to be predicted reaches.

In an alternative embodiment, the historical behavior information includes: the historical bubbling behavior information, the historical issuing behavior information and the historical completion behavior information are one or more of information.

In an alternative embodiment, the method further comprises:

acquiring fourth characteristic values of the user to be predicted under a plurality of user attribute characteristics;

the inputting the first feature value, the second feature value and the third feature value into a position prediction model trained in advance to obtain a position prediction result of the user to be predicted in the future preset time period includes:

and inputting the first characteristic value, the second characteristic value, the third characteristic value and the fourth characteristic value into a position prediction model trained in advance, and obtaining a position prediction result of the user to be predicted within the future preset time period.

In an alternative embodiment, the user attribute features include one or more of:

the number of orders, whether business people are present, whether tourist people are present, the area where the user's home is located, and the area where the user's company is located.

In an alternative embodiment, the position prediction model is trained using the following method:

acquiring first sample characteristic values of a plurality of sample users under at least one historical behavior characteristic, and second sample characteristic values and third sample characteristic values of each sample user under a target time characteristic and a target position characteristic in a historical preset time period; the sample users comprise positive sample users and negative sample users;

training the location prediction model based on the first, second, and third sample feature values of the respective sample users.

In an optional embodiment, the obtaining a first sample characteristic value of a plurality of sample users under at least one historical behavior characteristic, and a second sample characteristic value of each sample user under a target time characteristic and a third sample characteristic value under a target location characteristic within a historical preset time period includes:

for each sample area to be predicted in a plurality of sample areas to be predicted, acquiring a first sample characteristic value of a sample user corresponding to the sample area to be predicted under at least one historical behavior characteristic, and a second sample characteristic value and a third sample characteristic value of each sample user under a target time characteristic and a target position characteristic within a historical preset time period;

the training the location prediction model based on the first, second, and third sample feature values of the respective sample users comprises:

and training the position prediction model based on the first sample characteristic value, the second sample characteristic value and the third sample characteristic value of the sample user corresponding to each sample region to be predicted.

In an alternative embodiment, the training the position prediction model based on the first sample feature values and the second sample feature values of the respective sample users includes:

constructing a plurality of sub-decision trees based on the first sample characteristic value, the second sample characteristic value and the third sample characteristic value of each sample user;

determining a plurality of said sub-decision trees as said location prediction models.

In an optional implementation, constructing a plurality of sub-decision trees based on the first sample feature value, the second sample feature value and the third sample feature value of each sample user includes:

randomly determining a plurality of target characteristics from the historical behavior characteristics, the target time characteristics and the target area characteristics;

constructing a sub-decision tree of the current iteration period based on the characteristic values of the sample users under the target characteristics;

forming a current decision tree set based on the sub-decision trees of the current iteration period and the sub-decision trees of the historical iteration period, and determining the loss of the current decision tree set;

when the loss is larger than a preset loss threshold value, finishing a current iteration cycle, and returning to the historical behavior characteristics, the target time characteristics, the target area characteristics and the user attribute characteristics to randomly determine a plurality of target characteristics;

determining the current set of decision trees as the location prediction model if the penalty is not greater than a preset penalty threshold.

In an alternative embodiment, the determining the loss of the current decision tree set comprises:

inputting first test sample characteristic values of a plurality of test sample users under the historical behavior characteristics and second test sample characteristic values of the test sample users under to-be-predicted time characteristics and to-be-predicted area characteristics into the current decision tree set to obtain a position prediction result corresponding to each test sample user;

and determining the loss of the current decision tree set based on the position prediction result corresponding to each test sample user and the corresponding actual position.

In an alternative embodiment, the method further comprises: obtaining a fourth sample characteristic value of each sample user under a plurality of user attribute characteristics;

the location prediction model is trained based on the first sample feature value, the second sample feature value, the third sample feature value, and the fourth sample feature value of each sample user.

In an optional implementation manner, inputting the first feature value, the second feature value, and the third feature value into a position prediction model trained in advance, and obtaining a position prediction result of the user to be predicted in the future preset time period includes:

inputting the first characteristic value, the second characteristic value and the third characteristic value into a position prediction model trained in advance, and acquiring sub-prediction results of each sub-decision tree;

and carrying out weighted summation on the sub-prediction results of each sub-decision tree, and determining the position prediction result of the user to be predicted in the future time period.

In a second aspect, an embodiment of the present application provides a position prediction apparatus, including:

the system comprises an acquisition module, a prediction module and a prediction module, wherein the acquisition module is used for acquiring a first characteristic value of a user to be predicted under at least one historical behavior characteristic, and a second characteristic value and a third characteristic value of the user to be predicted under a target time characteristic and a target position characteristic in a future preset time period;

and the prediction module is used for inputting the first characteristic value, the second characteristic value and the third characteristic value into a position prediction model trained in advance and acquiring a position prediction result of the user to be predicted in the future preset time period.

In an alternative embodiment, the obtaining module is configured to obtain a first feature value of the user to be predicted under at least one historical behavior feature by:

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

the prediction module is used for acquiring a position prediction result of the user to be predicted in the future preset time period by adopting the following mode:

In an alternative embodiment, the method further comprises: a training module for training the position prediction model using:

In an optional embodiment, the training module is configured to obtain first sample feature values of a plurality of sample users under at least one historical behavior feature, and a second sample feature value of each of the sample users under a target time feature and a third sample feature value of each of the sample users under a target location feature within a historical preset time period:

In an alternative embodiment, the training module is configured to train the location prediction model based on the first sample feature values and the second sample feature values of the respective sample users in the following manner:

In an optional implementation manner, the training module is configured to construct a plurality of sub-decision trees based on the first sample feature value, the second sample feature value, and the third sample feature value of each sample user in the following manner:

In an alternative embodiment, the training module is configured to determine the loss of the current set of decision trees by:

In an optional implementation manner, the training module is further configured to obtain a fourth sample feature value of each sample user under multiple user attribute features;

the training module is configured to train the location prediction model based on the first sample feature value, the second sample feature value, and the third sample feature value of each sample user in the following manner:

In an optional implementation manner, the prediction module is configured to input the first feature value, the second feature value, and the third feature value into a position prediction model trained in advance, and obtain a position prediction result of the user to be predicted in the future preset time period by:

In a third aspect, an embodiment of the present application further provides an electronic device, including: 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 electronic device is running, the processor executing the machine-readable instructions to perform the steps of the position prediction method according to any one of the first aspect.

In a fourth aspect, the present application further provides a computer-readable storage medium, on which a computer program is stored, where the computer program is executed by a processor to perform the steps of the location prediction method according to any one of the first aspect.

According to the method, a first characteristic value of a user to be predicted under at least one historical behavior characteristic is obtained, and a second characteristic value of the user to be predicted under a target time characteristic and a third characteristic value of the user to be predicted under a target position characteristic in a future preset time period are obtained; and inputting the first characteristic value, the second characteristic value and the third characteristic value into a position prediction model trained in advance, obtaining a position prediction result of the user to be predicted in the future preset time period, and determining the position of the user to be predicted in the future preset time period with higher accuracy, so that the position information of the user can be predicted in advance, resource allocation can be performed in advance, and a relevant service strategy can be configured.

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 block diagram of a system 100 in one scenario in which the location determination method of some embodiments of the present application is applied;

FIG. 2 illustrates a schematic diagram of exemplary hardware and software components of an electronic device 200 of a server 110, a service requester terminal 130, a service provider terminal 140, which may implement the concepts of the present application, according to some embodiments of the present application;

fig. 3 is a flowchart illustrating a location prediction method according to an embodiment of the present application;

fig. 4 is a flowchart illustrating a specific method for obtaining a first feature value of a user to be predicted under at least one historical behavior feature in a location prediction method provided in embodiments of the present application;

FIG. 5 is a flow chart illustrating a specific method for training a location prediction model in a location prediction method provided in embodiments of the present application;

fig. 6 is a flowchart illustrating a specific method for constructing a plurality of sub-decision trees based on a first sample feature value, a second sample feature value, and a third sample feature value in a location prediction method according to embodiments of the present application;

fig. 7 is a flowchart illustrating a location prediction method provided in the second embodiment of the present application;

fig. 8 is a schematic structural diagram illustrating a position prediction apparatus 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 utilize the present disclosure, a related introduction is made in connection with an exemplary application scenario, "net appointment travel scenario". 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 described primarily in the context of predicting net appointment passenger locations, it should be understood that this is merely one exemplary embodiment. The application can be applied to any other traffic type. For example, the present application may be applied to different transportation system environments, including terrestrial, marine, or airborne, among others, or any combination thereof. The vehicle of the transportation system may include a taxi, a private car, a windmill, a bus, a train, a bullet train, a high speed rail, a subway, a ship, an airplane, a spacecraft, a hot air balloon, or an unmanned vehicle, etc., or any combination thereof. The present application may also include any service system for providing a service to a user based on the internet, for example, a system for sending and/or receiving a courier, a service system for a business to a seller. Applications of the system or method of the present application may include web pages, plug-ins for browsers, client terminals, customization systems, internal analysis systems, or artificial intelligence robots, among others, or any combination thereof.

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.

One aspect of the present application relates to a location prediction method. The method can input a first characteristic value of a user to be predicted under at least one historical behavior characteristic, a second characteristic value of the user to be predicted under a target time characteristic and a third characteristic value of the user to be predicted under a target position characteristic into a position prediction model trained in advance, and predict whether the user to be predicted appears at a target position in a future preset time period, so that the position information of the user can be predicted in advance, resource allocation can be performed in advance, and related service strategies can be configured.

It is to be noted that the prediction of the user location may also be performed based on the statistics of the user to be predicted, but the prediction accuracy in this way is low, and the location prediction method provided by the present application may predict the location of the user to be predicted with higher accuracy.

Fig. 1 is a block diagram of a system 100 in one scenario in which a location prediction method according to some embodiments of the present application is applied. For example, the 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 regular service, or any combination thereof. The system 100 may include one or more of a server 110, a network 120, a service requester terminal 130, a service provider terminal 140, and a database 150, and the server 110 may include a processor 220 therein that performs operations of instructions. The position prediction method provided by the embodiment of the present application may be applied to the server 110 in the system 100, and specifically, the processor 220 may execute the relevant operation instructions.

Fig. 2 illustrates a schematic diagram of exemplary hardware and software components of an electronic device 200 that may implement the server 110 of the present concepts according to some embodiments of the present application. For example, the processor 220 may be used on the electronic device 200 and to perform the functions herein.

The electronic device 200 may be a general purpose computer or a special purpose computer, both of which may be used to implement the location prediction method of the present application. Although only a single computer is shown, for convenience, the functions described herein may be implemented in a distributed fashion across multiple similar platforms to balance processing loads.

For example, the electronic device 200 may include a network port 210 connected to a network, one or more processors 220 for executing program instructions, a communication bus 230, and a different form of storage medium 240, such as a disk, ROM, or RAM, or any combination thereof. Illustratively, the computer platform may also include program instructions stored in ROM, RAM, or other types of non-transitory storage media, or any combination thereof. The method of the present application may be implemented in accordance with these program instructions. The electronic device 200 also includes an Input/Output (I/O) interface 250 between the computer and other Input/Output devices (e.g., keyboard, display screen).

For ease of illustration, only one processor is depicted in the electronic device 200. However, it should be noted that the electronic device 200 in the present application may also comprise a plurality of processors, and thus the steps performed by one processor described in the present application may also be performed by a plurality of processors in combination or individually. For example, if the processor of the electronic device 200 executes steps a and B, it should be understood that steps a and B may also be executed by two different processors together or separately in one processor. For example, the processor performs step a and the second processor performs step B, or the processor and the second processor perform steps a and B together.

Example one

Fig. 3 is a schematic diagram illustrating a position prediction method according to an embodiment of the present application, including S301 to S302.

S301: the method comprises the steps of obtaining a first characteristic value of a user to be predicted under at least one historical behavior characteristic, and obtaining a second characteristic value of the user to be predicted under a target time characteristic and a third characteristic value of the user to be predicted under a target position characteristic in a future preset time period.

In a specific implementation, the users to be predicted can be all users of the online car booking platform; in addition, because the number of users of the network appointment platform is very large, generally in the number of tens of millions or even billions, the position of each user needs to be determined, and the calculation amount is very large, the users can be screened from all the users of the network appointment platform based on certain conditions for more pertinence.

For example, the positions of the users at the time of bubbling can be periodically calculated, and if the bubbling position of any user in the period belongs to at least two cities or preset areas, the user is determined as the user to be predicted.

For example, the departure place of each user for placing an order may be calculated periodically, and if the departure place of any user in the period belongs to at least two cities or at least two preset areas, the user is determined as the user to be predicted.

The following describes, through I, ii, and iii, the manners of acquiring a first feature value of a user under at least one historical behavior feature, a second feature value under a target time feature, and a third feature value under a target location feature within a future preset time period, respectively:

i: obtaining a first characteristic value of a user to be predicted under at least one historical behavior characteristic:

referring to fig. 4, an embodiment of the present application provides a specific method for obtaining a first feature value of a user to be predicted under at least one historical behavior feature, including:

s401: the method comprises the steps of extracting area historical behavior information of a user to be predicted in an area to be predicted from at least one type of historical behavior information of the user to be predicted, and extracting the area time historical behavior information of the user to be predicted in the area to be predicted in each historical time period in a plurality of historical time periods.

In a specific implementation, the historical behavior of the user to be predicted includes: at least one of a bubbling action, a billing action, and a finishing action. Each historical behavior corresponds to one type of historical behavior information.

The bubbling behavior is that the service request end enters a home page of service software of the network car-booking platform, and a travel starting point and a travel terminal of the service request end are sent to the network car-booking platform. And the network car booking platform receives the travel starting point and the travel end point sent by the service request end and confirms to monitor the bubbling behavior of the service request end.

After obtaining a trip starting point and a trip end point sent by a service request end, the network car booking platform determines an order estimated price for the trip of the service request end, and sends the order estimated price to the service request end. And after receiving the estimated order price, the service request terminal generates order information based on the triggering of the passenger and sends the order information to the network car booking platform. This behavior is the issuing behavior of the service requester.

The network taxi appointment platform generates an order after receiving order information sent by the service request terminal, pushes the order to a plurality of service providing terminals meeting requirements, and matches the service providing terminals for the service request terminal based on feedback of the service providing terminals.

And when the service is completed by the service provider matched with the service request terminal, service completion information is fed back to the network car appointment platform. And the network car appointment platform initiates payment to the service request terminal. And after the service request end completes payment, the service request end is considered to have the order completion behavior.

It should be noted that the act of issuing a single is not equivalent to the act of completing a single. The user behavior is considered to be order-complete behavior only when the order is completed and payment is complete. If the service is cancelled by any one of the service providing terminal, the service request terminal or the network appointment vehicle in the midway of the service process, the order completion behavior corresponding to the order does not occur any more.

Each historical behavior corresponds to corresponding historical behavior information: historical behavior information corresponding to the bubbling behavior comprises historical bubbling behavior information; historical behavior information corresponding to the issuing behavior comprises historical issuing behavior information; the completion line information includes history completion line information. Each type of historical behavior information includes: historical behavior occurrence time and historical behavior occurrence place.

At least one kind of historical behavior information of the user is to be predicted, and the historical behavior information comprises at least one kind of historical behavior information of the three kinds of historical behaviors.

For example, at least one of historical behavior information of a user to be predicted, regional historical behavior information, and regional temporal historical behavior information may be obtained in the following manner.

(1) And acquiring historical behavior information X of the user to be predicted. Wherein, the ith piece of historical behavior information is recorded as xⁱ。

(2) According to the historical behavior occurrence place in each historical behavior information X, determining the historical behavior information X of the area to be predicted of the user to be predicted in the area to be predicted from the historical behavior information X_city. Wherein, the jth area historical behavior information is recorded as

(3) According to the historical behavior information X of the user to be predicted in the area to be predicted_cityDetermining the historical behavior information X of the user to be predicted at the area time of the area to be predicted in each historical time period in the plurality of historical time periods_city,t. Regional time historical behavior information X_city,tThere are multiple groups, the nth zone historical time behavior information of the mth group is recorded as

It should be noted that the historical behavior information X and the regional historical behavior information X_cityOne group is provided; regional time historical behavior information X_city,tThe same as the number of the history time period.

For example, the plurality of historical time periods may be specifically set according to actual needs, for example, 7 historical time periods may be determined according to the current time, and are respectively: 0-2 days, 2-5 days, 5-10 days, 10-20 days, 20-50 days, 50-100 days, more than 100 days. There are 7 sets of corresponding regional historical temporal behavior information.

S402: and determining a first characteristic value under the first historical behavior characteristic of the user to be predicted according to the at least one type of historical behavior information of the user to be predicted, the historical behavior information of the area and the historical behavior information of the area time.

S403: and determining a first characteristic value under a second historical behavior characteristic of the user to be predicted according to the at least one type of historical behavior information of the user to be predicted.

In a specific implementation, it is exemplary:

the first historical behavior characteristics include, but are not limited to, one or more of the following a 1-a 5:

a 1: the number Of occurrences Of the historical behaviors corresponding to each Of the plurality Of Point Of Interest (POI) classifications.

Here, the POI classification is a classification determined in advance for a different POI. For example: airports, entertainment, commerce, shopping, and the like.

a2, the number of occurrences of historical behavior per day in a preset time period.

Such as the number of occurrences of historical behavior on a monday through sunday, per day.

a3, the number of occurrences of historical behavior on weekdays. For example, historical behavior includes: the bubbling behavior and the ordering behavior are the times of the historical behaviors occurring on the working day, which are the times of the bubbling behavior and the ordering behavior occurring on the working day.

a4, the number of occurrences of historical behavior on non-working days.

a5, and the occurrence frequency corresponding to different historical behaviors. For example, if the historical behaviors include bubbling behavior and billing behavior, the occurrence times corresponding to different historical behaviors respectively include the occurrence times corresponding to the bubbling behavior and the occurrence times corresponding to the billing behavior.

The second historical behavior characteristics include, but are not limited to, one or more of the following b 1-b 5:

b 1: whether the area where the last historical behavior occurs is an area to be predicted or not;

b 2: the time interval between the time when the last historical behavior occurs and the time to be predicted;

b 3: POI classification of destination of last historical behavior;

b 4: POI classification of origin of last historical behavior

b 5: the number of areas reached by the user to be predicted.

Here, the area to be predicted and the time to be predicted are set according to reality.

Generally, the area to be predicted is an area that the user to be predicted has arrived at once. The time to be predicted is a future preset time period. The position prediction method provided by the embodiment of the application is to predict the probability that the user to be predicted appears in the area to be predicted in the future preset time period.

The first characteristic value of the user to be predicted under at least one historical behavior characteristic consists of three parts, including:

a first part: and determining a characteristic value under the first historical behavior characteristic corresponding to the at least one historical behavior information X of the user to be predicted according to the at least one historical behavior information X of the user to be predicted.

A second part, according to the area historical behavior information X of the area to be predicted in each historical time period in a plurality of historical time periods_cityDetermining the regional historical behavioral information X of the AND_cityCorresponding feature values under the first historical behavior feature.

The third part is used for predicting the regional time historical behavior information X of the region to be predicted according to the user to be predicted in each historical time period in a plurality of historical time periods_city,tDetermining historical time behavior information X of each group of regions_city,tA feature value under the first historical behavior feature.

II: acquiring a second characteristic value of the user to be predicted under the target time characteristic in a future preset time period:

here, the future preset time period is the time to be predicted.

The target time characteristics include: the future preset time period is a day of the week, and the future preset time period is a holiday or a weekday.

After the future preset time period with the predicted user is determined, according to the time information corresponding to the future preset time period, a second characteristic value of the user to be predicted under the target time characteristic in the future preset time period is determined.

III: acquiring a third characteristic value of the user to be predicted under the target position characteristic in a future preset time period;

here, the target position feature includes: attributes of the region to be predicted, for example: whether it is a tourist city, etc.

After the area to be predicted is determined, the corresponding third characteristic value can be determined according to the attribute of the area to be predicted and the attribute of the area to be predicted.

After receiving the above S301, after obtaining the first feature value, the second feature value, and the third feature value of the user to be predicted, the position prediction method provided in the embodiment of the present application further includes:

s302: and inputting the first characteristic value, the second characteristic value and the third characteristic value into a position prediction model trained in advance, and obtaining a position prediction result of the user to be predicted in the future preset time period.

In specific implementation, the obtained position prediction result of the user to be predicted in the future preset time period is to obtain the probability of the user to be predicted appearing in the position to be predicted in the future preset time period for prediction.

In addition, in another embodiment of the application, the probability that the user to be predicted appears at a plurality of positions to be predicted within a future preset time period can be predicted. In the prediction process, besides the first characteristic value and the second characteristic value of the user to be predicted, a corresponding third characteristic value of the user to be predicted for each area to be predicted is obtained, and the input of a plurality of groups of position prediction models is formed and is respectively input into the position prediction models, so that the probability of the user to be predicted appearing in each area to be predicted in a future preset time period is obtained.

For example, the probabilities of the respective occurrence of the user to be predicted in A, B, C three areas to be predicted within twenty-four hours of the future are determined.

The method comprises the steps of obtaining a first characteristic value a of a user to be predicted, a second characteristic value b under a target time characteristic in a future preset time period, and obtaining third characteristic values C1, C2 and C3 of the user to be predicted, wherein the third characteristic values C1, C2 and C3 correspond to regions A, B and C to be predicted respectively in the future preset time period.

Thus, three sets of input data are constructed, respectively: a. b and c 1; a. b and c 2; a. b and c 3.

And sequentially inputting the three groups of input data into a position prediction model trained in advance, and acquiring the probability of the occurrence of the user to be predicted at A, B and C within twenty-four hours in the future.

In another embodiment, after the probability of the user to be predicted appearing in the plurality of areas to be predicted is determined, the most probable position of the user to be predicted within a future preset time period can be determined according to the probability of the user to be predicted appearing in the plurality of areas to be predicted, so that resource scheduling and configuration of related service strategies can be performed in advance based on the prediction result.

Referring to fig. 5, an embodiment of the present application further provides a specific method for training a position prediction model, including:

s501: acquiring first sample characteristic values of a plurality of sample users under at least one historical behavior characteristic, and second sample characteristic values and third sample characteristic values of each sample user under a target time characteristic and a target position characteristic in a historical preset time period; the sample users include positive sample users and negative sample users.

S502: training the location prediction model based on the first, second, and third sample feature values of the respective sample users.

In a specific implementation, the historical preset time periods of different sample users may be the same or different. For example, the sample user a determines "11/15/2018" as the history preset time period, and the sample user B determines "10/2018" as the history preset time period. But the time length of the historical preset time period of different sample users is the same and is the same as the time length of the future preset time period.

And if the sample user appears in the area to be predicted within the historical preset time period, the sample user is a positive sample user. And if the sample user does not appear in the area to be predicted within the historical preset time period, the sample user is a negative sample user.

The probability that the positive sample user appears in the area to be predicted in the historical preset time period is 1 because the positive sample user appears in the area to be predicted in the historical preset time period; the negative sample user does not appear in the area to be predicted in the history preset time period, so that the probability that the negative sample user appears in the area to be predicted in the history preset time period is 0.

And in the process of training the model, after the first sample characteristic value, the second sample characteristic value and the third sample characteristic value of each sample user are input into the model position prediction model, the result output by the position prediction model is as close as possible to the probability of the position prediction model appearing in the region to be predicted.

The position prediction model may only predict the probability of occurrence of a certain to-be-predicted region in a future preset time period for the to-be-predicted user, or may predict the probability of occurrence of a plurality of to-be-predicted regions in the future preset time period for the to-be-predicted user, so that the determined sample user conditions are different for the two different conditions.

Firstly, the method comprises the following steps: when the position prediction model can only predict the probability of the user to be predicted appearing in a determined region to be predicted within the future preset limit, the following method can be adopted to determine the sample user:

and screening users appearing in the area to be predicted from the user database to serve as sample users.

For this case, the first sample feature value of the sample user may be determined using the following manner:

for each sample user, extracting the area historical behavior information of the sample user in the area to be predicted from at least one type of historical behavior information of the sample user, and extracting the area time historical behavior information of the sample user in the area to be predicted in each historical time period in a plurality of historical time periods;

determining a first sample characteristic value under a first historical behavior characteristic of the sample user according to the at least one type of historical behavior information, the regional historical behavior information and the regional time historical behavior information of the sample user;

and determining a first sample characteristic value under a second historical behavior characteristic of the sample user according to the at least one type of historical behavior information of the sample user.

Here, the first sample characteristic value of the sample user under the first historical behavior characteristic and the first sample characteristic value of the sample user under the second historical behavior characteristic are similar to the determination method of the first characteristic value of the user to be predicted, and are not described herein again.

It should be noted that the first sample characteristic value of the sample user is constructed based on the historical behavior information before the historical preset time period.

II, secondly: when the position prediction model can predict the probability of the user to be predicted appearing in a plurality of areas to be preset in a future preset time period, the sample user can be determined in the following way:

and screening users appearing in each area to be predicted from the user database to serve as sample users corresponding to each area to be predicted.

The first sample characteristic value, the second sample characteristic value and the third sample characteristic value of the sample user can be obtained by adopting the following modes:

specifically, the first sample feature value of each sample user may be obtained in any one of the following manners (1) and (2):

(1) for each area to be predicted and each sample user of the area to be predicted, extracting area historical behavior information of the sample user in the area to be predicted from at least one type of historical behavior information of the sample user, and extracting area time historical behavior information of the sample user in each historical time period in a plurality of historical time periods in the area to be predicted;

In this way, the sample users corresponding to each region to be predicted are screened based on the plurality of regions to be predicted.

(2) For a plurality of users, extracting the area historical behavior information of the user in at least one area to be predicted from at least one type of historical behavior information of the user, and extracting the area time historical behavior information of the user corresponding to each area to be predicted in at least one area to be predicted in each historical time period in a plurality of historical time periods.

According to the regional historical behavior information of a user in at least one region to be predicted and the regional time historical behavior information, corresponding to each region to be predicted, of the user in each historical time period of a plurality of historical time periods, determining a first sample characteristic value of the user in the at least one region to be predicted under a first historical behavior characteristic corresponding to each region to be predicted;

and determining a first sample characteristic value of the user in at least one region to be predicted and a second sample characteristic value of each region to be predicted corresponding to the second historical behavior characteristic according to the at least one type of historical behavior information of the sample user.

For example, user C has appeared in both areas M1 and M2 to be predicted.

From the historical behavior information of the user C, it is possible to extract the area historical behavior information corresponding to the area to be predicted M1, and extract the area time historical behavior information corresponding to the area to be predicted M1.

And then constructing a first sample characteristic value of the user C corresponding to the area to be predicted M1 according to the area historical behavior information corresponding to the area to be predicted M1 and the area time historical behavior information corresponding to the area to be predicted M1.

Furthermore, from the historical behavior information of the user C, it is also possible to extract the area historical behavior information corresponding to the area to be predicted M2 and extract the area time historical behavior information corresponding to the area with prediction M2. And then constructing a first sample characteristic value of the user C corresponding to the area to be predicted M2 according to the area historical behavior information corresponding to the area to be predicted M2 and the area time historical behavior information corresponding to the area to be predicted M2.

It should be noted that the sample users of different regions to be predicted may be the same or different. For example, if a user a appears in both the a area and the B area, the user a may be a sample user corresponding to the a area or a sample user corresponding to the B area.

However, since the regional historical behavior information of the first person in the region a is different from that of the second person in the region B, when the first person is a sample user of the region a and the second person in the region B, the first sample characteristic value corresponding to the region a and the first sample characteristic value corresponding to the region B are different. The corresponding third sample characteristic values are also different.

Meanwhile, in this case, the position prediction model may be trained in the following manner:

The embodiment of the present application further provides a specific method for training the position prediction model based on the first sample feature value and the second sample feature value of each sample user, including:

Specifically, referring to fig. 6, in this embodiment of the present application, a plurality of sub-decision trees are constructed based on a first sample feature value, a second sample feature value, and a third sample feature value in the following manner:

s601: randomly determining a plurality of target characteristics from the historical behavior characteristics, the target time characteristics and the target area characteristics;

s602: constructing a sub-decision tree of the current iteration period based on the characteristic values of the sample users under the target characteristics;

s603: forming a current decision tree set based on the sub-decision trees of the current iteration period and the sub-decision trees of the historical iteration period, and determining the loss of the current decision tree set;

s604: detecting whether the loss of the current decision tree set is greater than a preset loss threshold value or not; if yes, jumping to S601; the current iteration cycle is complete. If not, then jump to S605.

S605: determining the current set of decision trees as the location prediction model.

Specifically, a plurality of target features are randomly determined from the historical behavior features, the target time features and the target area features, and the position prediction model is trained according to the randomly determined plurality of target features, the feature value of each sample user under the corresponding target feature and the probability of the sample user appearing at the position to be predicted to obtain the decision tree set.

Then, inputting first test sample characteristic values of a plurality of test sample users under the historical behavior characteristics and second test sample characteristic values of the test sample users under to-be-predicted time characteristics and to-be-predicted area characteristics into the current decision tree set to obtain a position prediction result corresponding to each test sample user; and determining the loss of the current decision tree set based on the position prediction result corresponding to each test sample user and the corresponding actual position.

And adjusting target characteristics according to the loss of the decision tree set, and training the selection probability prediction model according to the new re-determined target characteristics to obtain the current decision tree set.

And then, verifying the current decision tree set based on the test sample user, and determining the loss of the current decision tree set.

And according to the steps, continuously iterating, optimizing and selecting the probability prediction model, and finally obtaining a decision tree set comprising N sub-decision trees through iteration, wherein N is a positive integer.

When the position prediction model predicts the position of a user to be predicted, the first characteristic value, the second characteristic value and the third characteristic value are input into a position prediction model trained in advance, and a sub-prediction result of each sub-decision tree is obtained;

According to the method, a first characteristic value of a user to be predicted under at least one historical behavior characteristic is obtained, and a second characteristic value of the user to be predicted under a target time characteristic and a third characteristic value of the user to be predicted under a target position characteristic in a future preset time period are obtained; and inputting the first characteristic value, the second characteristic value and the third characteristic value into a position prediction model trained in advance, obtaining a position prediction result of the user to be predicted in the future preset time period, and determining the position of the user to be predicted in the future preset time period with higher accuracy.

Example two

Fig. 7 illustrates another location prediction method provided in an embodiment of the present application, where the method includes:

s701: the method comprises the steps of obtaining a first characteristic value of a user to be predicted under at least one historical behavior characteristic, and obtaining a second characteristic value of the user to be predicted under a target time characteristic and a third characteristic value of the user to be predicted under a target position characteristic in a future preset time period.

The implementation manner of S701 is similar to that of S301, and is not described herein again.

S702: and acquiring fourth characteristic values of the user to be predicted under a plurality of user attribute characteristics.

Here, the user attribute features include one or more of:

S703: inputting the first characteristic value, the second characteristic value, the third characteristic value and the fourth characteristic value into a position prediction model trained in advance, and obtaining a position prediction result of the user to be predicted within the future preset time period.

Wherein, the implementation manner of S703 is similar to that of S302 described above. And will not be described in detail herein.

It should be noted that, in the second embodiment, during the training of the position prediction model, a fourth sample feature value of each sample user under multiple user attribute features is also obtained; when a position prediction model is trained, the position prediction model is trained based on the first sample feature value, the second sample feature value, the third sample feature value, and the fourth sample feature value of each sample user.

The specific training method is similar to the model training method in S302, and is not described herein again.

EXAMPLE III

Referring to fig. 8, an embodiment of the present application further provides a position prediction apparatus, including:

the obtaining module 81 is configured to obtain a first feature value of a user to be predicted under at least one historical behavior feature, and a second feature value of the user to be predicted under a target time feature and a third feature value of the user to be predicted under a target position feature within a future preset time period;

and the prediction module 82 is configured to input the first feature value, the second feature value, and the third feature value into a position prediction model trained in advance, and obtain a position prediction result of the user to be predicted in the future preset time period.

In a possible embodiment, the obtaining module 81 is configured to obtain a first feature value of the user to be predicted under at least one historical behavior feature by:

In one possible embodiment, the first historical behavior feature includes one or more of:

In one possible embodiment, the second historical behavior feature comprises one or more of:

In one possible embodiment, the historical behavior information includes: the historical bubbling behavior information, the historical issuing behavior information and the historical completion behavior information are one or more of information.

In a possible implementation, the obtaining module 81 is further configured to:

the prediction module 82 is configured to obtain a position prediction result of the user to be predicted in the future preset time period by using the following method:

In one possible embodiment, the user attribute feature includes one or more of the following:

In a possible embodiment, the method further comprises: a training module 83 configured to train the position prediction model using:

In a possible embodiment, the training module 83 is configured to obtain first sample feature values of a plurality of sample users under at least one historical behavior feature, and a second sample feature value of each sample user under a target time feature and a third sample feature value of each sample user under a target location feature within a historical preset time period:

the training module 83 is configured to train the location prediction model based on the first sample feature value, the second sample feature value, and the third sample feature value of each sample user in the following manner:

In a possible embodiment, the training module 83 is configured to train the location prediction model based on the first sample feature values and the second sample feature values of the respective sample users in the following manner:

In a possible implementation, the training module 83 is configured to construct a plurality of sub-decision trees based on the first sample feature value, the second sample feature value, and the third sample feature value of each sample user in the following manner:

In one possible embodiment, the training module 83 is configured to determine the loss of the current decision tree set by:

In a possible implementation manner, the training module 83 is further configured to obtain a fourth sample feature value of each sample user under multiple user attribute features;

In a possible implementation manner, the prediction module 82 is configured to input the first feature value, the second feature value, and the third feature value into a position prediction model trained in advance, and obtain a position prediction result of the user to be predicted in the future preset time period by:

The wired connections may include connections in the form of L AN, WAN, Bluetooth, ZigBee, or NFC, or the like, or any combination thereof.

As shown in fig. 2, an embodiment of the present application further provides an electronic device, including: a processor 220, a storage medium and a bus 230, wherein the storage medium stores machine-readable instructions executable by the processor 220, when the electronic device runs, the processor 220 communicates with the storage medium through the bus 230, and the processor 220 executes the machine-readable instructions to execute the steps of the position prediction method provided by the embodiment of the present application.

Embodiments of the present application further provide 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 computer program performs the steps of the position prediction method provided in the embodiments of the present application.

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 location prediction, the method comprising:

2. The method of claim 1, wherein the obtaining a first feature value of the user to be predicted under at least one historical behavior feature comprises:

3. The method of claim 2, wherein the first historical behavior feature comprises one or more of:

4. The method of claim 2, wherein the second historical behavior characteristic comprises one or more of:

5. The method of claim 2, wherein the historical behavior information comprises: the historical bubbling behavior information, the historical issuing behavior information and the historical completion behavior information are one or more of information.

6. The method of claim 1, further comprising:

7. The method of claim 6, wherein the user attribute features comprise one or more of:

8. The method of claim 1, wherein the location prediction model is trained using the following method:

9. The method of claim 8, wherein the obtaining of the first sample characteristic value of the plurality of sample users under at least one historical behavior characteristic, and the second sample characteristic value of each sample user under a target time characteristic and the third sample characteristic value under a target position characteristic within a historical preset time period comprises:

10. The method of claim 8, wherein training the location prediction model based on the first sample feature values and the second sample feature values of the respective sample users comprises:

11. The method according to claim 10, wherein constructing a plurality of sub-decision trees based on the first sample feature value, the second sample feature value and the third sample feature value of each sample user comprises:

12. The method of claim 11, wherein determining the loss of the current set of decision trees comprises:

13. The method of claim 8, further comprising: obtaining a fourth sample characteristic value of each sample user under a plurality of user attribute characteristics;

14. The method according to claim 10, wherein inputting the first feature value, the second feature value and the third feature value into a position prediction model trained in advance to obtain a position prediction result of the user to be predicted in the future preset time period comprises:

15. A position prediction apparatus, characterized in that the apparatus comprises:

16. The apparatus of claim 15, wherein the obtaining module is configured to obtain the first feature value of the user to be predicted under at least one historical behavior feature by:

17. The apparatus of claim 16, wherein the first historical behavior feature comprises one or more of:

18. The apparatus of claim 16, wherein the second historical behavior characteristic comprises one or more of:

19. The apparatus of claim 16, wherein the historical behavior information comprises: the historical bubbling behavior information, the historical issuing behavior information and the historical completion behavior information are one or more of information.

20. The apparatus of claim 15, wherein the obtaining module is further configured to:

21. The apparatus of claim 20, wherein the user attribute features comprise one or more of:

22. The apparatus of claim 15, further comprising: a training module for training the position prediction model using:

23. The apparatus of claim 22, wherein the training module is configured to obtain a first sample feature value of a plurality of sample users under at least one historical behavior feature, and a second sample feature value of each of the sample users under a target time feature and a third sample feature value of each of the sample users under a target location feature within a historical preset time period:

24. The apparatus of claim 22, wherein the training module is configured to train the location prediction model based on the first sample feature values and the second sample feature values of the respective sample users by:

25. The apparatus of claim 24, wherein the training module is configured to construct a plurality of sub-decision trees based on the first sample feature value, the second sample feature value, and the third sample feature value of each sample user by:

26. The apparatus of claim 25, wherein the training module is configured to determine the loss of the current set of decision trees by:

27. The apparatus of claim 22, wherein the training module is further configured to obtain a fourth sample feature value of each sample user under a plurality of user attribute features;

28. The apparatus of claim 24, wherein the prediction module is configured to input the first feature value, the second feature value, and the third feature value into a position prediction model trained in advance, and obtain a position prediction result of the user to be predicted in the future preset time period by:

29. An electronic 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 electronic device is operating, the processor executing the machine-readable instructions to perform the steps of the position prediction method according to any one of claims 1 to 14.

30. A computer-readable storage medium, having stored thereon a computer program for performing, when being executed by a processor, the steps of the position prediction method according to any one of claims 1 to 14.