CN112182431A - Reference point recommendation method, device, equipment and storage medium - Google Patents
Reference point recommendation method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN112182431A CN112182431A CN202011012047.1A CN202011012047A CN112182431A CN 112182431 A CN112182431 A CN 112182431A CN 202011012047 A CN202011012047 A CN 202011012047A CN 112182431 A CN112182431 A CN 112182431A
- Authority
- CN
- China
- Prior art keywords
- target
- current position
- user terminal
- reference location
- recommended
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9537—Spatial or temporal dependent retrieval, e.g. spatiotemporal queries
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/067—Enterprise or organisation modelling
Abstract
The application provides a reference place recommendation method, a reference place recommendation device, reference place recommendation equipment and a storage medium, wherein the method comprises the following steps: acquiring a reference location acquisition request initiated by a target user terminal, wherein the reference location acquisition request carries the current position of the target user terminal; determining a plurality of target reference positions which are within a preset range from the current position; inputting the identification of the target area to which the current position belongs and the identification of each of the plurality of target reference positions into a pre-trained vector generation model to obtain vector representation of the target area to which the current position belongs and vector representation of each of the plurality of target reference positions; the vector generation model is obtained by training a preset model by taking the incidence relation between areas with various sizes and each reference location as a training sample; and determining the reference location to be recommended to the target user terminal according to the similarity between the vector representation of the target area to which the current position belongs and the respective vector representations of the plurality of target reference locations.
Description
Technical Field
The present application relates to the field of information processing technologies, and in particular, to a reference location recommendation method, apparatus, device, and storage medium.
Background
In the existing network taxi appointment scene, a taxi getting-on place needs to be recommended for a user according to a taxi using order issued by the user. Wherein, recommending the boarding place for the user is generally realized through a sequencing model. The establishment of the sequencing model mainly depends on the statistical characteristics of the historical vehicle orders, such as order heat, vehicle getting-on heat and the like. However, the vehicle orders are distributed unevenly in different geographic areas, and the statistical characteristics cannot fully reflect the association correspondence between the user and the boarding place in the order issuing area, so that the accuracy of the boarding place recommended to the user is low.
Disclosure of Invention
In order to solve the above problems, the present application provides a reference location recommendation method, apparatus, device and storage medium, which aim to improve the accuracy of a boarding location recommended to a user.
In a first aspect of the embodiments of the present application, a reference location recommendation method is provided, where the method includes:
acquiring a reference location acquisition request initiated by a target user terminal, wherein the reference location acquisition request carries the current position of the target user terminal;
determining a plurality of target reference positions within a preset range away from the current position;
inputting the identifier of the target area to which the current position belongs and the identifiers of the target reference positions into a pre-trained vector generation model to obtain the vector representation of the target area to which the current position belongs and the vector representations of the target reference positions; the vector generation model is obtained by training a preset model by taking incidence relations between areas with various sizes and various reference points as training samples;
determining a reference location to be recommended according to the similarity between the vector representation of the target area to which the current position belongs and the respective vector representation of the plurality of target reference locations;
and pushing the reference place to be recommended to the target user terminal.
Optionally, the method further comprises:
obtaining a plurality of historical records, wherein each historical record comprises a historical position of a user terminal and a corresponding historical reference place;
determining areas of various sizes to which each historical position belongs;
according to the plurality of historical records, taking the region of each size and each reference point as nodes, constructing a weight network corresponding to the size, wherein the weight of the edge in the weight network represents the association degree between two nodes connected with the edge;
generating training samples according to the weight networks corresponding to the various sizes;
and training a preset model by using the training sample to obtain the vector generation model.
Optionally, generating a training sample according to a weight network corresponding to each of the multiple sizes, including:
obtaining a walking sequence taking any node as a starting point according to the weight network corresponding to each size;
generating a positive sample pair and a negative sample pair according to the hop count separated between the nodes aiming at each wandering sequence;
and dividing each positive sample pair and each negative sample generated based on the nodes corresponding to the mutually overlapped areas into a group of training samples to obtain a plurality of groups of training samples.
Optionally, the method further comprises:
acquiring a plurality of newly added records, wherein each newly added record comprises an actual position of a user terminal and a corresponding actual reference location;
generating a newly added training sample according to the newly added records;
updating the vector generation model according to the newly added training sample;
and inputting the identifier of the target area to which the current position belongs and the identifiers of the target reference positions into the updated vector generation model to obtain the updated vector representation of the target area to which the current position belongs and the updated vector representations of the target reference positions.
Optionally, the reference location obtaining request is a boarding point obtaining request; pushing the reference place to be recommended to the target user terminal, including:
and pushing the reference place to be recommended to the target user terminal as a boarding point to be recommended, and sending prompt information to the target user terminal to prompt a user of the target user terminal to move to the boarding point to be recommended.
In a second aspect of the embodiments of the present application, there is provided a reference location recommendation apparatus, including:
the system comprises a request acquisition module, a processing module and a processing module, wherein the request acquisition module is used for acquiring a reference location acquisition request initiated by a target user terminal, and the reference location acquisition request carries the current position of the target user terminal;
the first determination module is used for determining a plurality of target reference positions which are within a preset range away from the current position;
an obtaining module, configured to input an identifier of a target area to which the current position belongs and identifiers of the multiple target reference locations into a pre-trained vector generation model, so as to obtain a vector representation of the target area to which the current position belongs and vector representations of the multiple target reference locations; the vector generation model is obtained by training a preset model by taking incidence relations between areas with various sizes and various reference points as training samples;
a second determining module, configured to determine a reference location to be recommended according to similarities between the vector representations of the target area to which the current position belongs and the respective vector representations of the multiple target reference locations;
and the sending module is used for pushing the reference place to be recommended to the target user terminal.
In a third aspect of the embodiments of the present application, there is provided an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor implements the reference location recommendation method according to the first aspect when executed.
In a fourth aspect of the embodiments of the present application, a non-transitory computer-readable storage medium is provided, in which instructions are executable by a processor to perform operations performed by the reference location recommendation method according to any one of the first aspect.
By adopting the technical scheme of the embodiment of the application, the obtaining request of the reference location carrying the current position of the target user terminal, which is initiated by the target user terminal, can be obtained, and a plurality of target reference locations within a preset range away from the current position are determined; inputting the identification of the target area to which the current position belongs and the identification of each of the plurality of target reference positions into a pre-trained vector generation model to obtain vector representation of the target area to which the current position belongs and vector representation of each of the plurality of target reference positions; and then, determining the reference location to be recommended according to the similarity between the vector representation of the target area to which the current position belongs and the respective vector representation of the plurality of target reference locations.
The vector generation model is obtained by training the preset model by taking the incidence relations between the areas with various sizes and each reference point as training samples, so that the vector generation model can effectively construct the incidence relations between the target areas with different spatial scales and the reference location, for example, the incidence relations between the geographic areas with three spatial scales, namely large, medium and small, and the reference location are effectively described. Furthermore, the association degree between the target area to which the current position belongs and each reference location can be comprehensively depicted according to the vector representation of the target area generated by the vector generation model and the vector representation of each of the plurality of target reference locations, so that the accuracy of recommending the reference location to the user in the target area is improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required to be used in the description of the embodiments or the related art will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without inventive labor.
FIG. 1 is a flow chart illustrating the steps of a training vector generation model according to an embodiment of the present application;
FIG. 2 is a schematic diagram of a region partition according to an embodiment of the present application;
FIG. 3 is a diagram illustrating a weight network derived from one of the sizes according to an embodiment of the present application;
fig. 4 is a schematic diagram of a walk sequence corresponding to a weight network according to an embodiment of the present application;
FIG. 5 is a flow chart illustrating a process for updating a vector generation model according to an embodiment of the present application;
FIG. 6 is a flowchart illustrating steps of a method for recommending reference places according to an embodiment of the present application;
fig. 7 is a schematic structural diagram of a reference location recommendation apparatus in an embodiment of the present application.
Detailed Description
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 is obvious that the described embodiments are some, but not all, embodiments of the present application. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
It should be noted that the terms "first," "second," and the like in the description and claims of this application and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the application described herein are capable of operation in sequences other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.
In the related art, the place to pick up the car is recommended to the user, which generally depends on the statistical characteristics of the historical car order, such as order heat, car-picking heat, and the like. However, in this way, on one hand, the order distribution of the used cars is not uniform for different geographic areas, and the relationship between the distance between the departure area and the boarding point and the walking cost of the passengers cannot be completely reflected. On the other hand, when the statistical characteristics of the historical vehicle orders are taken into consideration, the statistical characteristics are not counted from a plurality of spatial scales, but are plotted from a single spatial scale, and therefore, the variation in the plotting tends to occur. Therefore, the two reasons for the above two aspects result in that the boarding place recommended to the user in the related art is not accurate enough.
Under the actual taxi taking scene, the taxi-boarding place recommendation is slightly influenced by people portrait (namely the preference of people) and is greatly influenced by geographical portrait, and on the basis, in order to improve the precision of the taxi-boarding place recommended by users, the inventor provides the following technical conception: according to historical orders, graph network link relations of the order issuing areas and the boarding places in various scales are built, sequence sets obtained by a plurality of graph network wandering are fused, association relations between the order issuing areas and the boarding places in various scales are built, then the association degrees of the current order issuing areas and the surrounding boarding places of the currently received order for the vehicle can be determined based on the established association relations, and the boarding places with higher association degrees with the current order issuing areas are determined according to the association degrees.
By adopting the technical concept of the application, the relationship between the user crowd preference and the geographic area can be better described, so that a more accurate boarding place is recommended to the user.
In the embodiment of the present application, the boarding location may also be referred to as a reference location, and the current invoice area may also be referred to as a target area.
Firstly, in order to improve the intelligence of determining a boarding place and the generalization capability of the method, a graph network link relation between an order issuing area and the boarding place can be constructed under various spatial scales according to historical orders, and a sequence set obtained by a plurality of graph network wandering is fused; based on a skip-gram mode, joint training is carried out on the walking sequences under multiple scales, so that a vector generation model is trained, and the vector generation model can depict the vector representation of the current order issuing area and the incidence relation between multiple boarding places through vector representation.
Referring to fig. 1, a flowchart of steps of training a vector generation model in the embodiment of the present application is shown, and as shown in fig. 1, a specific training process is as follows:
step S101: a plurality of history records are obtained, and each history record comprises the history position of one user terminal and a corresponding history reference place.
In this embodiment, the history records may be obtained by the server, and the plurality of history records may be records of the user's vehicle usage in a certain period of time before the current time. Each history record can comprise a history position and a history reference place of the user's issue. The historical position refers to the geographical position where the user places the order, and the historical reference place can be a boarding place recommended to the user.
In the embodiment of the application, the historical position issued by the user can be used for helping to determine the area where the user terminal is located at the moment, and the historical reference location can be used for being associated with the area where the user terminal is located at the moment.
Step S102: and determining various sizes of areas to which each historical position belongs.
In this embodiment, the plurality of sizes may be set in advance to sizes different from each other in size, and thus, for each of the historical positions, the historical position may be assigned to an area of a different size.
In practice, the same region may be divided according to a plurality of preset sizes, so as to obtain the regions of the region obtained under the division of the plurality of different sizes. The region is divided in multiple scales, and the region can be divided by adopting a geohash.
Illustratively, referring to fig. 2, which shows a schematic diagram of region division, in fig. 2, for a region a (which may refer to a city or a province), the region a may be divided in advance according to different sizes, so that for the historical location H, the historical location H may be located in the region a1 divided according to a small size, in the region a2 divided according to a medium size, and in the region A3 divided according to a large size.
After the areas with different sizes to which each historical position belongs are obtained, the association relationship between the geographical areas with different sizes and the boarding places can be conveniently and subsequently depicted.
Step S103: and constructing a weight network corresponding to each size by taking the region of each size and each historical reference point as nodes according to the plurality of historical records, wherein the weight of the edge in the weight network represents the association degree between two nodes connected with the edge.
In this embodiment, a node in the weight network may be connected to other nodes adjacent to the node, an edge formed by the connection of the node may have a weight, and the weight may represent a degree of association between two adjacent connected nodes, that is, the degree of association may refer to a degree of association between history reference locations. For example, if the probabilities of getting on the bus by the users in the same area are not very different, the association degree between the two nodes may be higher.
In specific implementation, for an area of one size, the area can be walked from one of the historical reference locations to another historical reference location, and when the area is walked to another historical reference location, a connecting line (i.e., an edge) formed by the historical reference location and the another historical reference location carries a weight, so that a weight network formed by a plurality of the historical reference locations is obtained. As described above, the weight value characterizes the degree of association between the two, wherein in one example, the weight value can be determined based on the number of connections between the two historical reference locations in the history.
In this embodiment, the number of the constructed weight networks is consistent with the number of the types of the adopted sizes, for example, if the geography is divided by adopting three sizes, there are three weight networks, and the three weight networks are weighted walking networks among the historical reference sites under the geographical area division of different sizes.
As shown in fig. 3, a schematic diagram of a weight network derived from one of the sizes is shown. Wherein V represents any one of the historical reference locations.
Step S104: and generating training samples according to the weight networks corresponding to the various sizes.
In this embodiment, after obtaining the multiple weight networks, for each weight network, a corresponding walking sequence may be obtained according to the Node2Vec method, so as to obtain walking sequences corresponding to the multiple weight networks, and further, based on the skip gram method, positive and negative sample pairs may be generated for the walking sequences List _1, List _2, and List _3, respectively.
In an example, in a specific implementation, a walking sequence using any node as a starting point may be obtained according to a weight network corresponding to each size; generating a positive sample pair and a negative sample pair according to the hop count separated between the nodes aiming at each wandering sequence; and dividing each positive sample pair and each negative sample generated based on the nodes corresponding to the mutually overlapped areas into a group of training samples to obtain a plurality of groups of training samples.
In this example, the hop count of the node may reflect whether the node is an adjacent node, for example, if the node 1 to the node2 are separated by 1 hop count, the node 1 and the node2 are adjacent nodes, and if the node 1 to the node2 are separated by 2 hop counts, the node 1 and the node2 are non-adjacent nodes. The positive sample pair refers to a sample pair formed by two adjacent nodes, and the negative sample pair refers to a sample pair formed by two non-adjacent nodes. The areas which are overlapped with each other refer to the areas which are overlapped by the areas corresponding to different sizes, so that the association relationship between the areas with different sizes and the historical reference nodes is described by dividing each positive sample pair and each negative sample pair generated by the nodes corresponding to the areas which are overlapped with each other into a group of training samples, and thus, model training is performed for the training samples by using the association relationship between the areas with different sizes and the historical reference nodes.
Each positive sample pair and each negative sample pair include an identifier of a node and an identifier of an area, where the identifier of the node may be an ID of the node, specifically, an ID of a reference location, and the identifier of the area may be an ID of the area.
In this embodiment, referring to fig. 4, a schematic diagram of a walking sequence corresponding to one of the weight networks is shown, where "v" in fig. 4 represents a jump-starting node, where fig. 4 shows a walking sequence diagram corresponding to each of the weight networks with four sizes, circles in the diagram show one node, and arrows show a jump sequence.
Step S105: and training a preset model by using the training sample to obtain the vector generation model.
In this embodiment, because multiple sets of training samples are obtained, a preset model may be trained using the multiple sets of training samples, where the preset model may be configured to output vector representations corresponding to the regions and vector representations of the historical reference locations according to the input multiple sets of training samples. Wherein, the preset model may be a skip-gram model.
In this embodiment, because the multiple sets of training samples are based on samples generated in different scale regions, the node embedding can be learned in a joint training manner when the preset model is trained.
In an example, with the increase of the vehicle-using orders sent by the user, training samples can be generated according to the newly added vehicle-using orders, and the vector generation model is continuously updated, so that the precision of the vector generation model for generating the vector representation of each reference location is improved, and the precision of recommending the vehicle-getting-on location for the user is further improved.
Specifically, referring to fig. 5, a flowchart illustrating steps for updating the vector generation model is shown, which may specifically include the following steps:
step S201: and acquiring a plurality of newly added records, wherein each newly added record comprises the actual position of a user terminal and a corresponding actual reference position.
In this embodiment, the new record may refer to a car taking record corresponding to a car taking order newly received by the server after the training of the vector generation model is completed, and similarly, each new record may include an actual position and an actual reference location of the user terminal. The actual location may be the location where the user terminal was when the order for the car was sent.
Step S202: and generating a newly added training sample according to the newly added records.
In this embodiment, the process of generating the newly added training sample may be: constructing a newly added weight network corresponding to each size by taking the area of each size and each actual reference point as a node according to areas of various sizes to which each actual position in the newly added records belongs, wherein the weight of the edge in the newly added weight network represents the association degree between two nodes connected with the edge; and generating a new training sample according to the new weight value network corresponding to each of the multiple sizes.
The process of constructing the new weighting value network is similar to the process of step S103, and for relevant points, reference may be made to the description of step S103, which is not described herein again. The process of generating the additional training samples is similar to the process of step S104, and for the relevant points, reference may be made to the description of step S104.
Step S203: and updating the vector generation model according to the newly added training sample.
In this embodiment, updating the vector generation model may refer to: and training the vector generation model again by adding a new training sample, thereby obtaining a trained vector generation model, namely an updated vector generation model.
As the training samples are increased, the model training effect is better, and the timeliness of geographic features between the geographic region and the reference location can be reflected better, the updated vector generation model can output the vector representation of the reference location and the vector representation of the region more accurately.
According to the steps, after the vector generation model is trained, the vector generation model can construct the incidence relation between the ordering areas and the boarding places with different spatial scales, so that under the condition that a new vehicle order is generated, the vector generation model can output the vector representation of the target area and the vector representation of each of the target reference places, and the boarding place recommended to the target user can be determined according to the vector representation of the target area and the vector representation of each of the target reference places.
Referring to fig. 6, a flowchart illustrating steps of a reference location recommendation method according to an embodiment of the present application is shown, and as shown in fig. 6, the method may specifically include the following steps:
step S301: and acquiring a reference location acquisition request initiated by a target user terminal, wherein the reference location acquisition request carries the current position of the target user terminal.
In this embodiment, the reference location obtaining request may be a request sent by the user when the user needs to use the vehicle, and the reference location obtaining request may carry the current location of the target user terminal.
Step S302: and determining a plurality of target reference positions within a preset range from the current position.
In general, a plurality of target reference points, which may be points near a target user that may be a boarding point, are stored near the current location of the user terminal. In this embodiment, the preset range may be preset according to experience or requirements, for example, the preset range is set to 1 km, and then the reference location within a radius range less than 1 km from the current location of the target user may be determined as the target reference location.
Step S303: and inputting the identifier of the target area to which the current position belongs and the identifiers of the target reference positions into a pre-trained vector generation model to obtain the vector representation of the target area to which the current position belongs and the vector representations of the target reference positions.
The vector generation model is obtained by training a preset model by taking the incidence relation between areas with various sizes and each reference location as a training sample.
In this embodiment, a target area to which the current position belongs may be determined, and since a boarding point needs to be recommended for the user according to the current position of the user, in order to improve accuracy of recommending a boarding point for the user and reduce a boarding cost, an area to which the current position belongs in each area divided in the minimum size may be determined as the target area based on each area divided in the minimum size. The identification of the target area may refer to an ID of the target area, and the identification of the target reference location may also be an ID of the target reference location. In specific implementation, the identifier of the target region to which the current position belongs and the identifiers of the plurality of target reference locations may be input into a pre-trained vector generation model, so as to obtain a vector representation of the target region to which the current position belongs and vector representations of the plurality of target reference locations.
Step S304: and determining a reference location to be recommended according to the similarity between the vector representation of the target area to which the current position belongs and the respective vector representation of the plurality of target reference locations.
In this embodiment, since the vector representation of the target region to which the current position belongs and the respective vector representations of the plurality of target reference locations are obtained, the similarity between the vector representation of the target region and the plurality of target reference locations can be calculated respectively.
The similarity may reflect a degree of association between a target area to which the current position belongs and the target reference location, or may be understood as reflecting a cost of getting on the vehicle between the target area to which the current position belongs and the target reference location, where the higher the similarity is, the higher the degree of association between the target area and the target reference location is, the lower the cost of getting on the vehicle is, and the corresponding target reference location is a reference location that the user in the target area will select with a higher probability.
Therefore, the target reference location with the highest similarity or with the similarity higher than the preset similarity can be determined as the reference location to be recommended to the target user.
It should be noted that, in the embodiment of the present application, in a practical process, a newly added training sample is generated according to a plurality of newly added records, and then the vector generation model is updated by using the newly added training sample, so that when obtaining a vector representation of a target region to which a current position belongs and respective vector representations of a plurality of target reference locations, an identifier of the target region to which the current position belongs and respective identifiers of the plurality of target reference locations may be input into the updated vector generation model, so as to obtain an updated vector representation of the target region to which the current position belongs and respective updated vector representations of the plurality of target reference locations. And then, determining the reference location to be recommended according to the similarity between the updated vector representation of the target area to which the current position belongs and the respective updated vector representations of the plurality of target reference locations.
Step S305: and pushing the reference place to be recommended to the target user terminal.
In this embodiment, the determined reference places to be recommended may be pushed to the target user terminal, where when there are multiple reference places to be recommended, all the multiple reference places to be recommended may be recommended to the user, or a reference place to be recommended, which is closest to the current position of the target user or has the highest similarity with the target area, of the multiple reference places to be recommended may be pushed to the target user terminal.
In an example, the reference location obtaining request is a boarding point obtaining request, and when the reference location to be recommended is pushed to the target user terminal, the reference location to be recommended can be pushed to the target user terminal as the boarding point to be recommended, and prompt information is sent to the target user terminal to prompt a user of the target user terminal to travel to the boarding point to be recommended.
The get-on point acquiring request may be a get-on point recommending request sent by a user when the user needs to use a car, for example, the target user needs the platform to recommend a suitable get-on point to the target user. In this case, the reference location to be recommended may be a boarding point to be recommended, and when the boarding point to be recommended is pushed to the target user terminal, the reference location to be recommended may send prompt information to the target user terminal together, where the prompt information may be information that prompts the user of the target user terminal to travel to the boarding point to be recommended.
In one example, the prompt message may include route information from the current location of the target user terminal to the pick-up point to be recommended, so that the user of the target user terminal may be helped to travel to the pick-up point to be recommended according to the route information. In yet another example, the prompt message may include a live-action image of the pick-up point to be recommended, so that the user of the target user terminal may be helped to accurately locate the pick-up point to be recommended according to the live-action image. Of course, in yet another example, the prompt information may include the route information and the live-action image of the boarding point to be recommended, as described above.
By adopting the technical scheme of the application, the method has the following advantages:
firstly, according to historical orders, a graph network link relation between the order issuing areas with various scales and the boarding places is built, and a sequence set obtained by a plurality of graph network wanders is fused, so that an incidence relation between the order issuing areas with various scales and the boarding places is built, then, the similarity between the current order issuing area of the currently received order for the user and the surrounding boarding places can be determined based on the established incidence relation, and further, the relation between the order issuing area of the user and the boarding places is accurately drawn, so that the more accurate boarding places are obtained.
Secondly, the efficiency of determining the accurate boarding place can be improved by the technical scheme that the vector representation of the current issuing area and the respective vector representations of the plurality of boarding places are depicted through the vector generation model at the training position. In the process of using the vector generation model, the vector generation model can be continuously updated along with the increase of orders, so that the accuracy of recommending the boarding places can be further improved.
Based on the same inventive concept as the above embodiment, referring to fig. 7, a block diagram of a reference location recommendation apparatus of this embodiment is shown, and the apparatus may specifically include the following modules:
a request obtaining module 701, configured to obtain a reference location obtaining request initiated by a target user terminal, where the reference location obtaining request carries a current location of the target user terminal;
a first determining module 702, configured to determine a plurality of target reference locations within a preset range from the current position;
an obtaining module 703, configured to input an identifier of a target area to which the current position belongs and identifiers of the multiple target reference locations into a pre-trained vector generation model, so as to obtain a vector representation of the target area to which the current position belongs and vector representations of the multiple target reference locations; the vector generation model is obtained by training a preset model by taking incidence relations between areas with various sizes and various reference points as training samples;
a second determining module 704, configured to determine a reference location to be recommended according to similarities between the vector representations of the target area to which the current position belongs and the respective vector representations of the multiple target reference locations;
a sending module 705, configured to push the reference location to be recommended to the target user terminal.
Optionally, the apparatus may further include the following modules:
the history record obtaining module is used for obtaining a plurality of history records, and each history record comprises the history position of a user terminal and a corresponding history reference place;
the third determining module is used for determining areas of various sizes to which each historical position belongs;
a network construction module, configured to construct a weight network corresponding to each size according to the multiple history records, with the area of each size and each reference point as a node, where a weight of an edge in the weight network represents a degree of association between two nodes connected to the edge;
the sample generating module is used for generating training samples according to the weight networks corresponding to the various sizes;
and the training module is used for training a preset model by using the training sample to obtain the vector generation model.
Optionally, the sample generation module may specifically include the following units:
a sequence obtaining unit, configured to obtain a walking sequence using any node as a starting point according to a weight network corresponding to each size;
the positive and negative sample generating unit is used for generating a positive sample pair and a negative sample pair according to the hop count separated between the nodes aiming at each wandering sequence;
and the sample establishing unit is used for dividing each positive sample pair and each negative sample generated based on the nodes corresponding to the mutually overlapped areas into a group of training samples to obtain a plurality of groups of training samples.
Optionally, the apparatus may further include the following modules:
the system comprises a newly added record obtaining module, a newly added record obtaining module and a newly added recording obtaining module, wherein the newly added records are used for obtaining a plurality of newly added records, and each newly added record comprises the actual position of a user terminal and a corresponding actual reference place;
a newly added sample generation module for generating newly added training samples according to the newly added records;
the updating module is used for updating the vector generation model according to the newly added training sample;
an input module, configured to input the identifier of the target area to which the current position belongs and the identifiers of the multiple target reference locations into the updated vector generation model, so as to obtain an updated vector representation of the target area to which the current position belongs and updated vector representations of the multiple target reference locations.
Optionally, the reference location obtaining request is a boarding point obtaining request; the sending module may be specifically configured to push the reference location to be recommended to the target user terminal as a boarding point to be recommended, and send prompt information to the target user terminal to prompt a user of the target user terminal to travel to the boarding point to be recommended.
It should be noted that the device embodiments are similar to the method embodiments, so that the description is simple, and reference may be made to the method embodiments for relevant points.
The embodiment of the present application further provides an electronic device, which may be used to execute the video stream processing method and may include a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor is configured to execute the reference location recommendation method.
Embodiments of the present application also provide a non-transitory computer-readable storage medium, wherein instructions of the storage medium, when executed by a processor, enable the processor to perform operations performed to implement the reference location recommendation method described above.
The embodiments in the present specification are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other.
As will be appreciated by one of skill in the art, embodiments of the present application may be provided as a method, apparatus, or computer program product. Accordingly, embodiments of the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, embodiments of the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
Embodiments of the present application are described with reference to flowchart illustrations and/or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing terminal to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing terminal to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing terminal to cause a series of operational steps to be performed on the computer or other programmable terminal to produce a computer implemented process such that the instructions which execute on the computer or other programmable terminal provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
While preferred embodiments of the present application have been described, additional variations and modifications of these embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, it is intended that the appended claims be interpreted as including the preferred embodiment and all such alterations and modifications as fall within the true scope of the embodiments of the application.
Finally, it should also be noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or terminal. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or terminal that comprises the element.
The reference location recommendation method, apparatus, device and storage medium provided by the present application are introduced in detail, and a specific example is applied in the present application to explain the principle and implementation of the present application, and the description of the above embodiment is only used to help understand the method and core ideas of the present application; meanwhile, for a person skilled in the art, according to the idea of the present application, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present application.
Claims (8)
1. A method for reference location recommendation, the method comprising:
acquiring a reference location acquisition request initiated by a target user terminal, wherein the reference location acquisition request carries the current position of the target user terminal;
determining a plurality of target reference positions within a preset range away from the current position;
inputting the identifier of the target area to which the current position belongs and the identifiers of the target reference positions into a pre-trained vector generation model to obtain the vector representation of the target area to which the current position belongs and the vector representations of the target reference positions; the vector generation model is obtained by training a preset model by taking the incidence relation between areas with various sizes and each reference location as a training sample;
determining a reference location to be recommended according to the similarity between the vector representation of the target area to which the current position belongs and the respective vector representation of the plurality of target reference locations;
and pushing the reference place to be recommended to the target user terminal.
2. The method of claim 1, further comprising:
obtaining a plurality of historical records, wherein each historical record comprises a historical position of a user terminal and a corresponding historical reference place;
determining areas of various sizes to which each historical position belongs;
according to the plurality of historical records, taking the region of each size and each historical reference point as nodes, constructing a weight network corresponding to the size, wherein the weight of the edge in the weight network represents the association degree between two nodes connected with the edge;
generating training samples according to the weight networks corresponding to the various sizes;
and training a preset model by using the training sample to obtain the vector generation model.
3. The method of claim 2, wherein generating training samples according to the weight networks corresponding to the plurality of sizes comprises:
obtaining a walking sequence taking any node as a starting point according to the weight network corresponding to each size;
generating a positive sample pair and a negative sample pair according to the hop count separated between the nodes aiming at each wandering sequence;
and dividing each positive sample pair and each negative sample generated based on the nodes corresponding to the mutually overlapped areas into a group of training samples to obtain a plurality of groups of training samples.
4. The method according to any one of claims 1-3, further comprising:
acquiring a plurality of newly added records, wherein each newly added record comprises an actual position of a user terminal and a corresponding actual reference location;
generating a newly added training sample according to the newly added records;
updating the vector generation model according to the newly added training sample;
and inputting the identifier of the target area to which the current position belongs and the identifiers of the target reference positions into the updated vector generation model to obtain the updated vector representation of the target area to which the current position belongs and the updated vector representations of the target reference positions.
5. The method according to any one of claims 1 to 3, wherein the reference location acquisition request is a pick-up point acquisition request; pushing the reference place to be recommended to the target user terminal, including:
and pushing the reference place to be recommended to the target user terminal as a boarding point to be recommended, and sending prompt information to the target user terminal to prompt a user of the target user terminal to move to the boarding point to be recommended.
6. A reference location recommendation apparatus, the apparatus comprising:
the system comprises a request acquisition module, a processing module and a processing module, wherein the request acquisition module is used for acquiring a reference location acquisition request initiated by a target user terminal, and the reference location acquisition request carries the current position of the target user terminal;
the first determination module is used for determining a plurality of target reference positions which are within a preset range away from the current position;
an obtaining module, configured to input an identifier of a target area to which the current position belongs and identifiers of the multiple target reference locations into a pre-trained vector generation model, so as to obtain a vector representation of the target area to which the current position belongs and vector representations of the multiple target reference locations; the vector generation model is obtained by training a preset model by taking incidence relations between areas with various sizes and various reference points as training samples;
a second determining module, configured to determine a reference location to be recommended according to similarities between the vector representations of the target area to which the current position belongs and the respective vector representations of the multiple target reference locations;
and the sending module is used for pushing the reference place to be recommended to the target user terminal.
7. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor when executing implementing the reference location recommendation method of any one of claims 1-5.
8. A computer-readable storage medium storing a computer program for causing a processor to execute the reference location recommending method according to any one of claims 1 through 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011012047.1A CN112182431A (en) | 2020-09-22 | 2020-09-22 | Reference point recommendation method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011012047.1A CN112182431A (en) | 2020-09-22 | 2020-09-22 | Reference point recommendation method, device, equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112182431A true CN112182431A (en) | 2021-01-05 |
Family
ID=73956224
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011012047.1A Pending CN112182431A (en) | 2020-09-22 | 2020-09-22 | Reference point recommendation method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112182431A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113722614A (en) * | 2021-08-05 | 2021-11-30 | 上海钧正网络科技有限公司 | Method and device for determining getting-on position and server |
-
2020
- 2020-09-22 CN CN202011012047.1A patent/CN112182431A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113722614A (en) * | 2021-08-05 | 2021-11-30 | 上海钧正网络科技有限公司 | Method and device for determining getting-on position and server |
| CN113722614B (en) * | 2021-08-05 | 2024-05-07 | 上海钧正网络科技有限公司 | Method and device for determining boarding location and server |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110929162B (en) | Recommendation method and device based on interest points, computer equipment and storage medium | |
| CN102906540B (en) | Information providing device and information providing method | |
| Lai et al. | Urban traffic Coulomb’s law: A new approach for taxi route recommendation | |
| Teimoori et al. | A secure cloudlet-based charging station recommendation for electric vehicles empowered by federated learning | |
| CN110782065A (en) | Electric vehicle charging pile recommendation method, server, terminal and system | |
| CN111127130B (en) | Energy site recommendation method based on user preference, storage medium and electronic equipment | |
| CN109785611B (en) | Unmanned vehicle control method, device, server and storage medium | |
| CN109282825B (en) | Road marking method and device | |
| CN111739285A (en) | Traffic flow prediction method, device, equipment and computer storage medium | |
| CN111311193B (en) | Method and device for configuring public service resources | |
| CN113722409A (en) | Method and device for determining spatial relationship, computer equipment and storage medium | |
| CN110015095B (en) | Method and equipment for charging mobile equipment | |
| CN112182431A (en) | Reference point recommendation method, device, equipment and storage medium | |
| CN110830915A (en) | Method and device for determining starting point position | |
| CN111710153B (en) | Traffic flow prediction method, device, equipment and computer storage medium | |
| CN111159510A (en) | Bus stop inquiry method based on intelligent terminal, intelligent terminal and device | |
| CN116069794A (en) | Map information updating method, device, equipment and computer readable storage medium | |
| CN103927869A (en) | Public transportation information processing method and device | |
| CN112504291A (en) | Vehicle navigation method and device | |
| CN107656967A (en) | A kind of scene information processing method and processing device | |
| KR20220095452A (en) | Computer system of providing service consumption plan for efficient service migration in mobile edge computing environments, and method thereof | |
| CN112566043A (en) | MAC address identification method and device, storage medium and electronic equipment | |
| CN111405467A (en) | Vehicle communication method, vehicle-road cooperation system, related equipment and readable medium | |
| CN115691211B (en) | Parking stall recommendation method and device, vehicle and storage medium | |
| CN111815101B (en) | Information processing method and device, storage medium and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |