Data recommendation method and device, electronic equipment and computer readable medium

Abstract

The invention discloses a data recommendation method and device, electronic equipment and a computer readable medium, and relates to the technical field of computers.The specific implementation mode comprises the steps that in response to a data recommendation request, a corresponding road area identifier and a corresponding site identifier are obtained; acquiring a corresponding road area local feature according to the road area identifier; acquiring navigation distances between corresponding stations and between road areas in the stations according to the station identifiers, and further generating a distance matrix based on each navigation distance; constructing a road network graph based on the distance matrix, and converting the road network graph into a low-dimensional dense vector based on a graph embedding mode; and based on the road area local features and the low-dimensional dense vectors, predicting to obtain a target recommendation operation mode. According to the method, road network features are supplemented by introducing a graph embedding method, the spatial global / adjacent features of stations / road areas are expressed more completely, and the accuracy of recommending operation modes is improved.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a data recommendation method, apparatus, electronic device, and computer-readable medium. Background Technology

[0002] Currently, in modern express delivery companies, all express delivery services are completed by terminal delivery personnel through pickup and delivery. The efficiency and timeliness of these delivery personnel directly affect the company's revenue and development. Generally, a work area, called a route zone, is defined for each delivery person to ensure timely pickup and delivery within that zone. Existing algorithms partially consider the spatial characteristics of stations / route zones, i.e., the distance from the station to the route zone, which can lead to biases in the recommended work patterns. Summary of the Invention

[0003] In view of this, embodiments of this application provide a data recommendation method, apparatus, electronic device, and computer-readable medium, which can solve the problem that in existing express delivery companies, when recommending the work mode of delivery personnel, the spatial characteristics of the station / road area are partially considered, resulting in a bias in the recommendation of the work mode.

[0004] To achieve the above objectives, according to one aspect of an embodiment of this application, a data recommendation method is provided, comprising:

[0005] In response to a data recommendation request, obtain the corresponding road area identifier and station identifier;

[0006] Obtain the corresponding local features of the road area based on the road area markings;

[0007] Based on the station identifier, obtain the navigation distances between corresponding stations and between road areas within a station, and then generate a distance matrix based on each navigation distance;

[0008] A road network map is constructed based on the distance matrix, and then the road network map is converted into a low-dimensional dense vector based on graph embedding.

[0009] Based on local road area features and low-dimensional dense vectors, the target recommended operation mode is predicted.

[0010] Optionally, the navigation distance is the distance between stations and between road zones within a station that can be reached by the route.

[0011] Optionally, a distance matrix is ​​generated based on each navigation distance, including:

[0012] Fill the positions of the preset distance matrix placeholders with each navigation distance to generate the corresponding distance matrix.

[0013] Optionally, a road network map is constructed based on a distance matrix, including:

[0014] Obtain the individual data corresponding to each matrix point in the distance matrix;

[0015] A road network map is constructed based on each matrix point and its corresponding single-quantity data.

[0016] Optionally, the road network graph is converted into a low-dimensional dense vector based on graph embedding, including:

[0017] For each node in the road network graph, a corresponding low-dimensional dense node vector is obtained through node embedding.

[0018] Optionally, the target recommended job pattern is predicted, including:

[0019] Local road area features are word-embedded to transform them into local road area feature vectors;

[0020] The local feature vector of the road area is concatenated with a low-dimensional dense vector to generate a concatenated vector;

[0021] Based on the concatenated vectors, the target recommendation job pattern is predicted.

[0022] Optionally, based on the concatenated vector, the target recommendation job pattern is predicted, including:

[0023] The concatenated vector is input into the job pattern recommendation model to output the target recommended job pattern.

[0024] In addition, this application also provides a data recommendation device, including:

[0025] The first acquisition unit is configured to acquire the corresponding road area identifier and station identifier in response to a data recommendation request;

[0026] The second acquisition unit is configured to acquire the corresponding local features of the road area based on the road area identifier;

[0027] The distance matrix generation unit is configured to obtain the navigation distances between corresponding stations and between road zones within a station based on the station identifier, and then generate a distance matrix based on each navigation distance.

[0028] The transformation unit is configured to construct a road network map based on a distance matrix, and then convert the road network map into a low-dimensional dense vector based on graph embedding.

[0029] The prediction unit is configured to predict the target recommendation operation mode based on local road area features and low-dimensional dense vectors.

[0030] Optionally, the navigation distance is the distance between stations and between road zones within a station that can be reached by the route.

[0031] Optionally, the distance matrix generation unit is further configured to:

[0032] Fill the positions of the preset distance matrix placeholders with each navigation distance to generate the corresponding distance matrix.

[0033] Optionally, the conversion unit is further configured to:

[0034] Obtain the individual data corresponding to each matrix point in the distance matrix;

[0035] A road network map is constructed based on each matrix point and its corresponding single-quantity data.

[0036] Optionally, the conversion unit is further configured to:

[0037] For each node in the road network graph, a corresponding low-dimensional dense node vector is obtained through node embedding.

[0038] Optionally, the prediction unit is further configured to:

[0039] Local road area features are word-embedded to transform them into local road area feature vectors;

[0040] The local feature vector of the road area is concatenated with a low-dimensional dense vector to generate a concatenated vector;

[0041] Based on the concatenated vectors, the target recommendation job pattern is predicted.

[0042] Optionally, the prediction unit is further configured to:

[0043] The concatenated vector is input into the job pattern recommendation model to output the target recommended job pattern.

[0044] In addition, this application also provides a data recommendation electronic device, including: one or more processors; and a storage device for storing one or more programs, which, when executed by one or more processors, cause the one or more processors to implement the data recommendation method as described above.

[0045] In addition, this application also provides a computer-readable medium having a computer program stored thereon, which, when executed by a processor, implements the data recommendation method as described above.

[0046] One embodiment of the above invention has the following advantages or beneficial effects: In response to a data recommendation request, this application obtains the corresponding road area identifier and station identifier; obtains the corresponding local features of the road area based on the road area identifier; obtains the navigation distances between stations and between road areas within a station based on the station identifier, and then generates a distance matrix based on each navigation distance; constructs a road network map based on the distance matrix, and then converts the road network map into a low-dimensional dense vector using graph embedding; and predicts the target recommended operation mode based on the local features of the road area and the low-dimensional dense vector. By introducing graph embedding to supplement road network features, the spatial global / adjacent features of stations / road areas are expressed more completely, improving the accuracy of operation mode recommendations.

[0047] The further effects of the aforementioned unconventional alternative methods will be explained below in conjunction with specific implementation methods. Attached Figure Description

[0048] The accompanying drawings are provided to better understand this application and do not constitute an undue limitation thereof. Wherein:

[0049] Figure 1 This is a schematic diagram of the main flow of a data recommendation method provided according to an embodiment of this application;

[0050] Figure 2 This is a schematic diagram of the main flow of a data recommendation method provided according to an embodiment of this application;

[0051] Figure 3 This is a schematic diagram of a road network map structure for a data recommendation method provided according to an embodiment of this application;

[0052] Figure 4 This is a schematic diagram illustrating an application scenario of a data recommendation method provided according to an embodiment of this application;

[0053] Figure 5 This is a schematic diagram of the main units of a data recommendation device according to an embodiment of this application;

[0054] Figure 6 This is an exemplary system architecture diagram to which embodiments of this application can be applied;

[0055] Figure 7 This is a schematic diagram of the structure of a computer system suitable for implementing terminal devices or servers in the embodiments of this application. Detailed Implementation

[0056] The following description, in conjunction with the accompanying drawings, illustrates exemplary embodiments of this application, including various details to aid understanding. These should be considered merely exemplary. Therefore, those skilled in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this application. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description. It should be noted that the collection, gathering, updating, analysis, processing, use, transmission, and storage of user personal information involved in the technical solutions of this disclosure all comply with relevant laws and regulations, are used for legitimate purposes, and do not violate public order and good morals. Necessary measures are taken to prevent unauthorized access to user personal information data and to safeguard user personal information security, network security, and national security.

[0057] Figure 1 This is a schematic diagram of the main flow of a data recommendation method provided according to an embodiment of this application, as shown below. Figure 1 As shown, data recommendation methods include:

[0058] Step S101: In response to the data recommendation request, obtain the corresponding road area identifier and station identifier.

[0059] In this embodiment, the execution entity of the data recommendation method (e.g., a server) can detect in real time whether there is a data recommendation request via a wired or wireless connection. For example, the data recommendation request could be a request to recommend work modes to delivery personnel or taxi drivers; this embodiment does not specifically limit the content of the data recommendation request. After detecting a data recommendation request, the execution entity can obtain the corresponding road area identifier and station identifier. This embodiment uses recommending work modes to delivery personnel as an example to illustrate the specific implementation of the data recommendation method. For example, the road area identifier can be used to represent the code or name corresponding to the work area (i.e., road area) designated for delivery personnel; this embodiment does not specifically limit the road area identifier. For example, the station identifier can be used to represent the code or name of the logistics distribution station connected to each road area; this embodiment does not specifically limit the station identifier. Figure 3 As shown, a road zone can be a triangular shape, and a station can be a circular shape. A station can connect one or more road zones, and stations can also be connected to each other.

[0060] Step S102: Obtain the corresponding local features of the road area based on the road area identifier.

[0061] like Figure 4As shown, local features of a road area may include on-time performance, order volume, labor efficiency, and the number of POIs, etc. This application embodiment does not specifically limit the local features of the road area. Specifically, on-time performance = on-time order volume / total order volume; labor efficiency = total order volume / number of people / time; and the number of POIs refers to the number of points of interest within the road area. POI is an abbreviation for "Point of Interest." In a geographic information system, a POI can be a building, a shop, a mailbox, a bus stop, etc.

[0062] The executing entity can obtain local features of the corresponding road area based on the road area identifier. For example, it can obtain the on-time rate, number of orders, manpower efficiency, and number of POIs within the corresponding road area.

[0063] Step S103: Obtain the navigation distances between corresponding stations and between road areas within a station based on the station identifier, and then generate a distance matrix based on each navigation distance.

[0064] Specifically, navigation distance refers to the distances between accessible routes and between road areas within a station. Navigation distance includes the reachable distance between stations and the reachable distance between road areas within a station. Reachable distance differs from straight-line distance and ensures accessibility between stations and road areas within a station.

[0065] A distance matrix is ​​a square matrix (two-dimensional array) containing a set of distances between pairs of elements. Depending on the application, the distances used to define this matrix may or may not be a metric. If there are N elements (i.e., N nodes), the matrix is ​​of size N×N. For example, the elements (i.e., the nodes in the distance matrix) can indicate navigation distances.

[0066] A distance matrix is ​​obtained by arranging navigation distances according to a preset layout. This application embodiment does not specifically limit the preset layout. The distance matrix can include both reachable distances between stations and reachable distances between road sections within a station; that is, it can simultaneously include reachable distances between stations and reachable distances between road sections within a station.

[0067] Step S104: Construct a road network map based on the distance matrix, and then convert the road network map into a low-dimensional dense vector based on graph embedding.

[0068] Each node in the distance matrix represents a navigation distance. Based on the arrangement of nodes in the distance matrix and the corresponding navigation distances, a road network map is constructed. For example, a road network map could be like this: Figure 3As shown, the road network map structure (considering only the edges connecting stations and omitting the connections between road areas under different stations, conforming to real-world delivery rules) can be represented by different graphics for different node types in the road network map. Node types can include road areas and stations. The executing entity can convert each element in the distance matrix (i.e., the node in the distance matrix) into two corresponding nodes in the road network map and an edge connecting the two nodes. The length L of the edge corresponds to the navigation distance L' of the corresponding element in the distance matrix (i.e., the node in the distance matrix). It can be understood that L can be equal to L', or it can be a preset scaled-down or scaled-up version of L'. This application embodiment does not specifically limit the relationship between the length of the edge in the road network map and the corresponding navigation distance.

[0069] After constructing the road network map, the implementing entity can convert each node in the road network map into a low-dimensional dense vector based on node embedding in graph embedding. Graph embedding maps the original graph data (usually a sparse high-dimensional matrix) into a low-dimensional dense vector. Node embedding in graph embedding can be used to obtain a vector representation for each node in the road network map through embedding.

[0070] Furthermore, as one implementation of this application, the executing entity can also embed the entire road network map into a vector based on the whole map embedding method in graph embedding. This vector can then be used to determine whether two communities containing stations and road areas are similar, providing a broader range of similarity comparison methods at the community level. This can be used to expand community business, enabling a single data recommendation calculation to be applied to multiple highly similar communities, thereby improving business processing efficiency and the accuracy of data recommendation.

[0071] Step S105: Based on local road area features and low-dimensional dense vectors, the target recommended operation mode is predicted.

[0072] The executing entity can convert local road area features into local road area feature vectors, and then input them together with the obtained low-dimensional dense vectors into the operation mode recommendation model to output the corresponding target recommendation mode. For example, the target recommendation mode can be group, shift, dedicated collection, or two-way relay. Group refers to delivery personnel forming a group to divide tasks into time slots. Shift refers to dividing delivery personnel into multiple shifts to take turns performing pickup and delivery business. Dedicated collection refers to focusing solely on pickup without delivery. Two-way relay refers to setting up a relay point, where dedicated personnel relay delivery, and others deliver to the relay point. This application embodiment does not specifically limit the target recommendation mode.

[0073] This embodiment responds to data recommendation requests by obtaining corresponding road area identifiers and station identifiers; obtaining corresponding local features of the road area based on the road area identifiers; obtaining navigation distances between stations and between road areas within a station based on the station identifiers, and then generating a distance matrix based on each navigation distance; constructing a road network map based on the distance matrix, and then converting the road network map into a low-dimensional dense vector using graph embedding; and predicting the target recommended operation mode based on the local features of the road area and the low-dimensional dense vector. By introducing graph embedding to supplement road network features, the spatial global / adjacent features of stations / road areas are expressed more completely, improving the accuracy of operation mode recommendations.

[0074] Figure 2 This is a schematic diagram of the main flow of a data recommendation method provided according to an embodiment of this application, such as... Figure 2 As shown, data recommendation methods include:

[0075] Step S201: In response to the data recommendation request, obtain the corresponding road area identifier and station identifier.

[0076] Step S202: Obtain the corresponding local features of the road area based on the road area identifier.

[0077] Step S203: Obtain the navigation distances between corresponding stations and between road areas within a station based on the station identifier, and then generate a distance matrix based on each navigation distance.

[0078] Specifically, generating a distance matrix based on each navigation distance includes filling each navigation distance into the positions of preset distance matrix placeholders to generate the corresponding distance matrix.

[0079] Step S204: Construct a road network map based on the distance matrix, and obtain the corresponding low-dimensional dense node vector for each node in the road network map through node embedding.

[0080] Specifically, constructing a road network map based on a distance matrix includes: obtaining the individual quantity data corresponding to each matrix point in the distance matrix; and constructing a road network map based on each matrix point and its corresponding individual quantity data.

[0081] The executing entity can obtain the paired road area and station corresponding to each matrix point (i.e., element, or node) in the distance matrix, or the paired station and its corresponding single-volume data. For example, when a matrix point W in the distance matrix corresponds to road area A and station A', the executing entity can obtain the single-volume data Q of road area A and the single-volume data Q' corresponding to station A'. The single-volume data Q and single-volume data Q' are determined as the single-volume data corresponding to a matrix node W in the obtained distance matrix. Each graph node in the constructed road network map contains the single-volume data corresponding to that node. The single-volume data may include road area single-volume, pickup single-volume, or delivery single-volume, etc., and this application embodiment does not specifically limit the single-volume data.

[0082] Step S205: Word embedding is performed on the local features of the road area to transform them into local feature vectors of the road area.

[0083] The implementing entity can convert the timeliness rate, number of transactions, per capita efficiency, and number of POIs in the local features of the road area into corresponding vectors, so as to facilitate subsequent vector splicing and fusion operations and improve the speed and accuracy of data recommendation.

[0084] Step S206: The local feature vector of the road area is concatenated with the low-dimensional dense vector to generate a concatenated vector.

[0085] The executing entity can add the local feature vector of the road area to the low-dimensional dense vector to obtain the concatenated vector after vector addition.

[0086] Step S207: Based on the concatenated vector, predict the target recommendation job mode.

[0087] Specifically, based on the concatenated vector, the target recommended job pattern is predicted, including: inputting the concatenated vector into the job pattern recommendation model to output the target recommended job pattern.

[0088] This application's embodiments supplement road network features by introducing graph embedding, which more completely expresses the spatial global / adjacent features of stations / road areas, thereby improving the accuracy of recommending work modes.

[0089] Figure 4 This is a schematic diagram illustrating an application scenario of a data recommendation method provided according to an embodiment of this application. Figure 4In this framework, on-time performance = on-time orders / total orders; labor efficiency = total orders / number of people / time; and the number of Points of Interest (POIs) refers to the number of points of interest within a road area. POI is an abbreviation for "Point of Interest." In a Geographic Information System (GIS), a POI can be a building, a shop, a mailbox, a bus stop, etc. A distance matrix is ​​calculated using the coordinates of the station / road area center point, including navigation distances within a station and between stations. These distances differ from straight-line distances; they are navigation distances to ensure reachability. A road network map is constructed using the distance matrix and order volume information. The node2vec method is used to calculate the graph embedding for each station / road area. The graph embedding results are then added to the features of the original method (i.e., previously isolated features such as on-time performance, order volume, labor efficiency, and the number of POIs), and concatenated to form new features. These new features are then input into XGBoost for model training to obtain the job mode recommendation model. In the model prediction step, the station / road area map embedding is also input along with other features (such as on-time rate, order volume, labor efficiency, and POI quantity, etc., local features of the road area) to perform classification prediction, complete the work mode classification, and make work mode recommendations.

[0090] Among them, node2vec is a graph embedding computation method that comprehensively considers depth-first search and breadth-first search of neighborhood. The principle of the node2vec algorithm is briefly described as follows: node2vec is a model used to generate node vectors in a network. The input is the network structure (which can be unweighted), and the output is the vector of each node. For each node, the goal is to maximize the probability of its neighboring nodes when representing the node as a vector, expressed in logarithmic form. The probability of a node vector appearing in its neighboring nodes is equal to the product of the probabilities of each of its neighboring nodes. The probability of each other node under the given node vector can be represented by softmax. These probabilities are then combined to obtain the objective function. Given preset values ​​and corresponding neighboring nodes, the objective function can be obtained by maximizing the objective function, thus yielding the vector of each node.

[0091] This application uses XGBOOST, an optimized distributed gradient boosting library designed for high efficiency, flexibility, and portability. XGBOOST (eXtreme Gradient Boosting) is an algorithm based on GBDT. Gradient Boosting Decision Tree (GBDT) is an additive model based on the boosting ensemble idea. During training, it uses a forward distribution algorithm for greedy learning, learning a CART tree in each iteration to fit the residuals between the predictions of the previous t-1 trees and the true values ​​of the training samples. The work mode is the model selection for terminal delivery personnel in the logistics industry. By grouping / segmenting the delivery personnel's work, work efficiency can be improved. There are four work modes: small group / shift / dedicated / two-way connection. Different work modes need to be set according to different work areas, and specific recommendations can be made based on the geographical attributes within the work area. This application's embodiment uses graph embedding to characterize different road areas, which can effectively preserve and compress spatiotemporal features for calculation, improving the accuracy of recommendations. Compared to previous methods that relied solely on isolated features such as local quantities and areas, graph embedding more comprehensively carries global and neighboring features, improving the accuracy of data recommendation. This application's embodiment introduces graph embedding of stations / road areas as supplementary features. Compared to the conventional use of local road area features, this more completely expresses the spatial global / neighboring features of stations / road areas, thus improving the accuracy of work pattern recommendations.

[0092] Figure 5 This is a schematic diagram of the main units of a data recommendation device according to an embodiment of this application. For example... Figure 5 As shown, the data recommendation device 500 includes a first acquisition unit 501, a second acquisition unit 502, a distance matrix generation unit 503, a conversion unit 504, and a prediction unit 505.

[0093] The first acquisition unit 501 is configured to acquire the corresponding road area identifier and station identifier in response to a data recommendation request.

[0094] The second acquisition unit 502 is configured to acquire the corresponding local features of the road area based on the road area identifier.

[0095] The distance matrix generation unit 503 is configured to obtain the navigation distances between corresponding stations and between road areas within a station based on the station identifier, and then generate a distance matrix based on each navigation distance.

[0096] Transformation unit 504 is configured to construct a road network map based on a distance matrix, and then convert the road network map into a low-dimensional dense vector based on graph embedding.

[0097] Prediction unit 505 is configured to predict the target recommendation operation mode based on local road area features and low-dimensional dense vectors.

[0098] In some embodiments, navigation distance is the distance between stations and between road zones within a station that can be reached.

[0099] In some embodiments, the distance matrix generation unit 503 is further configured to fill each navigation distance into the position of a preset distance matrix placeholder to generate the corresponding distance matrix.

[0100] In some embodiments, the conversion unit 504 is further configured to: acquire the unit data corresponding to each matrix point in the distance matrix; and construct a road network map based on each matrix point and its corresponding unit data.

[0101] In some embodiments, the transformation unit 504 is further configured to obtain a corresponding low-dimensional dense node vector for each node in the road network diagram through node embedding.

[0102] In some embodiments, the prediction unit 505 is further configured to: embed words into local road area features to transform them into local road area feature vectors; concatenate the local road area feature vectors with low-dimensional dense vectors to generate a concatenated vector; and predict the target recommendation operation mode based on the concatenated vector.

[0103] In some embodiments, the prediction unit 505 is further configured to input the concatenated vector into the job mode recommendation model to output the target recommended job mode.

[0104] It should be noted that the data recommendation method and data recommendation device in this application are related in terms of specific implementation, so repeated content will not be described again.

[0105] Figure 6 An exemplary system architecture 600 is shown that can be applied to the data recommendation method or data recommendation apparatus of the embodiments of this application.

[0106] like Figure 6 As shown, system architecture 600 may include terminal devices 601, 602, and 603, a network 604, and a server 605. Network 604 serves as the medium for providing communication links between terminal devices 601, 602, and 603 and server 605. Network 604 may include various connection types, such as wired or wireless communication links or fiber optic cables, etc.

[0107] Users can use terminal devices 601, 602, and 603 to interact with server 605 via network 604 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 601, 602, and 603, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social media platform software, etc. (for example only).

[0108] Terminal devices 601, 602, and 603 can be various electronic devices with data recommendation processing screens and support web browsing, including but not limited to smartphones, tablets, laptops, and desktop computers.

[0109] Server 605 can be a server providing various services, such as a backend management server supporting data recommendation requests submitted by users using terminal devices 601, 602, and 603 (this is just an example). The backend management server can respond to data recommendation requests by obtaining the corresponding road area identifier and station identifier; obtaining the corresponding local features of the road area based on the road area identifier; obtaining the navigation distances between stations and between road areas within a station based on the station identifier, and then generating a distance matrix based on these navigation distances; constructing a road network map based on the distance matrix, and then converting the road network map into a low-dimensional dense vector using graph embedding; and predicting the target recommended operation mode based on the local features of the road area and the low-dimensional dense vector. By introducing graph embedding to supplement road network features, the spatial global / adjacent features of stations / road areas are expressed more completely, improving the accuracy of operation mode recommendations.

[0110] It should be noted that the data recommendation method provided in this application embodiment is generally executed by server 605, and correspondingly, the data recommendation device is generally set in server 605.

[0111] It should be understood that Figure 6 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0112] The following is for reference. Figure 7 It shows a schematic diagram of the structure of a computer system 700 suitable for implementing a terminal device according to the embodiments of this application. Figure 7 The terminal device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0113] like Figure 7As shown, the computer system 700 includes a central processing unit (CPU) 701, which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) 702 or programs loaded from storage section 708 into random access memory (RAM) 703. The RAM 703 also stores various programs and data required for the operation of the computer system 700. The CPU 701, ROM 702, and RAM 703 are interconnected via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0114] The following components are connected to the I / O interface 705: an input section 706 including a keyboard, mouse, etc.; an output section 707 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 708 including a hard disk, etc.; and a communication section 709 including a network interface card such as a LAN card, modem, etc. The communication section 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to the I / O interface 705 as needed. A removable medium 711, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 710 as needed so that computer programs read from it can be installed into the storage section 708 as needed.

[0115] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 709, and / or installed from removable medium 711. When the computer program is executed by central processing unit (CPU) 701, it performs the functions defined above in the system of this application.

[0116] It should be noted that the computer-readable medium shown in this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. Computer-readable storage media can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0117] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0118] The units described in the embodiments of this application can be implemented in software or hardware. The described units can also be housed in a processor; for example, a processor can be described as including a first acquisition unit, a second acquisition unit, a distance matrix generation unit, a transformation unit, and a prediction unit. The names of these units do not necessarily limit the specific unit itself.

[0119] In another aspect, this application also provides a computer-readable medium, which may be included in the device described in the above embodiments; or it may exist independently and not assembled into the device. The computer-readable medium carries one or more programs that, when executed by the device, cause the device to respond to a data recommendation request by: obtaining corresponding road area identifiers and station identifiers; obtaining corresponding local features of the road area based on the road area identifiers; obtaining navigation distances between stations and between road areas within stations based on the station identifiers; generating a distance matrix based on each navigation distance; constructing a road network map based on the distance matrix; converting the road network map into a low-dimensional dense vector using graph embedding; and predicting the target recommendation operation mode based on the local features of the road area and the low-dimensional dense vector.

[0120] According to the technical solution of this application embodiment, by introducing a graph embedding method to supplement road network features, the spatial global / adjacent features of stations / road areas are expressed more completely, thereby improving the accuracy of recommending operation modes.

[0121] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A data recommendation method, characterized in that, include: In response to a data recommendation request, obtain the corresponding road area identifier and station identifier; Obtain the corresponding local features of the road area based on the road area identifier; Based on the station identifier, obtain the corresponding navigation distances between stations and between road areas within a station, and then generate a distance matrix based on each navigation distance; A road network map is constructed based on the distance matrix, and then the road network map is converted into a low-dimensional dense vector based on graph embedding. Based on the local features of the road area and the low-dimensional dense vector, the target recommended operation mode is predicted.

2. The method according to claim 1, characterized in that, The navigation distance refers to the distance between stations and between road zones within a station that can be reached.

3. The method according to claim 1, characterized in that, The generation of a distance matrix based on each of the navigation distances includes: The navigation distances are filled into the positions of the preset distance matrix placeholders to generate the corresponding distance matrix.

4. The method according to claim 1, characterized in that, The construction of the road network map based on the distance matrix includes: Obtain the single data corresponding to each matrix point in the distance matrix; A road network map is constructed based on each matrix point and its corresponding single-quantity data.

5. The method according to claim 1, characterized in that, The graph embedding-based method for converting the road network graph into a low-dimensional dense vector includes: For each node in the road network diagram, a corresponding low-dimensional dense node vector is obtained through node embedding.

6. The method according to claim 1, characterized in that, The prediction yields the target recommended job pattern, including: The local features of the road area are word-embedded to transform them into local feature vectors of the road area; The local feature vector of the road area is concatenated with the low-dimensional dense vector to generate a concatenated vector; Based on the concatenated vector, the target recommendation job mode is predicted.

7. The method according to claim 6, characterized in that, The step of predicting the target recommendation job mode based on the concatenated vector includes: The concatenated vector is input into the job mode recommendation model to output the target recommended job mode.

8. A data recommendation device, characterized in that, include: The first acquisition unit is configured to acquire the corresponding road area identifier and station identifier in response to a data recommendation request; The second acquisition unit is configured to acquire the corresponding local features of the road area based on the road area identifier; The distance matrix generation unit is configured to obtain the navigation distances between corresponding stations and between road zones within a station based on the station identifier, and then generate a distance matrix based on each of the navigation distances. The transformation unit is configured to construct a road network map based on the distance matrix, and then convert the road network map into a low-dimensional dense vector based on graph embedding. The prediction unit is configured to predict the target recommendation operation mode based on the local features of the road area and the low-dimensional dense vector.

9. The apparatus according to claim 8, characterized in that, The navigation distance refers to the distance between stations and between road zones within a station that can be reached.

10. The apparatus according to claim 8, characterized in that, The distance matrix generation unit is further configured to: The navigation distances are filled into the positions of the preset distance matrix placeholders to generate the corresponding distance matrix.

11. A data recommendation electronic device, characterized in that, include: One or more processors; Storage device for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-7.

12. A computer-readable medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1-7.

Images