CN114663165A - Site selection prediction method, site selection prediction device, site selection prediction apparatus, site selection prediction medium, and program product - Google Patents

Site selection prediction method, site selection prediction device, site selection prediction apparatus, site selection prediction medium, and program product Download PDF

Info

Publication number
CN114663165A
CN114663165A CN202210408033.4A CN202210408033A CN114663165A CN 114663165 A CN114663165 A CN 114663165A CN 202210408033 A CN202210408033 A CN 202210408033A CN 114663165 A CN114663165 A CN 114663165A
Authority
CN
China
Prior art keywords
data
site selection
graph
prediction
target area
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202210408033.4A
Other languages
Chinese (zh)
Inventor
吴思奥
暨光耀
张�浩
傅媛媛
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Industrial and Commercial Bank of China Ltd ICBC
Original Assignee
Industrial and Commercial Bank of China Ltd ICBC
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Industrial and Commercial Bank of China Ltd ICBC filed Critical Industrial and Commercial Bank of China Ltd ICBC
Priority to CN202210408033.4A priority Critical patent/CN114663165A/en
Publication of CN114663165A publication Critical patent/CN114663165A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0201—Market modelling; Market analysis; Collecting market data
    • G06Q30/0204—Market segmentation
    • G06Q30/0205—Location or geographical consideration
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • 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
    • G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
    • G06Q40/02—Banking, e.g. interest calculation or account maintenance

Abstract

The disclosure provides a site selection prediction method for a network point, which can be applied to the technical field of artificial intelligence or the financial field. The method comprises the following steps: acquiring signaling data of a target area; storing the signaling data of the target area into a graph database to construct a site selection information association graph; dividing the network point addressing information association diagram based on preset distance granularity to obtain m local network point addressing information association diagram data, wherein m is an integer greater than or equal to 2; and preprocessing the local site selection information association graph and inputting the preprocessed local site selection information association graph into a pre-trained site selection prediction model to obtain a site selection prediction result, wherein the site selection prediction result comprises a service grade prediction grade of a target area, and the pre-trained site selection prediction model is constructed on the basis of a graph convolution neural network. The present disclosure also provides a device, an apparatus, a storage medium and a program product for predicting site selection of a network node.

Description

Site selection prediction method, site selection prediction device, site selection prediction apparatus, site selection prediction medium, and program product
Technical Field
The present disclosure relates to the field of artificial intelligence techniques or the field of finance, and in particular, to a method, an apparatus, a device, a medium, and a program product for predicting site selection of a website.
Background
With the continuous development of social economy, the competition of banks is increasingly intense. The traditional physical bank outlets are one of important marketing channels and the most visual way for displaying the brand images of banks, and the address selection of the physical bank outlets is directly related to the development and economic benefits of the banks. The selection of a scientific and reasonable website address becomes an important condition for the banks to gain advantages in the competition of the financial market, and has important significance for the development of the banks and the improvement of the core competitiveness. Therefore, establishing a rapid, accurate and automatic bank branch site selection evaluation and recommendation method is very important for seizing regional financial markets and competing for customer resources. The existing bank website selection method mainly comprises two methods, namely a first manual site selection method and a second method for constructing a site selection model by using signaling data according to a specified weight formula or by establishing a logistic regression machine learning mode.
In the course of implementing the disclosed concept, the inventors found that there are at least the following problems in the prior art: the first manual site selection method is characterized in that data such as people flow and traffic are collected manually, then, a site selection model is established by experts through the data to select a site address, the time and the labor are consumed for site selection, the efficiency is low, many unreasonable places are possibly generated, the resource repetition and waste are caused, and the bank income is poor. The second method is to construct an addressing model by using signaling data according to a designated weight formula or by establishing a logistic regression machine learning mode, and solves the problem of high labor cost of traditional addressing to a certain extent. However, the designated weight formula is often set through human experience, and the influence of subjective judgment may cause the address selection result to be inaccurate. By establishing a simple model for logistic regression machine learning, the associated information between feature data cannot be fully mined, so that the accuracy of site selection of network points is also influenced.
Disclosure of Invention
In view of the foregoing, embodiments of the present disclosure provide a method, an apparatus, a device, a medium, and a program product for improving accuracy and scientificity of site selection prediction for a website.
According to a first aspect of the present disclosure, there is provided a mesh point addressing prediction method, including: acquiring signaling data of a target area; storing the signaling data of the target area into a graph database to construct a site selection information association graph; dividing the network point addressing information association diagram based on preset distance granularity to obtain m local network point addressing information association diagram data, wherein m is an integer greater than or equal to 2; and preprocessing the local site selection information association graph and inputting the preprocessed local site selection information association graph into a pre-trained site selection prediction model to obtain a site selection prediction result, wherein the site selection prediction result comprises a service grade prediction grade of a target area, and the pre-trained site selection prediction model is constructed on the basis of a graph convolution neural network.
According to an embodiment of the present disclosure, the signaling data includes at least two of economic data, geographical data, physical condition data, and peer competition data.
According to the embodiment of the disclosure, the building of the site selection information association diagram comprises the following steps: storing the attributes of a target area and the target area into a graph database by taking the target area as a core node and the attribute of the target area as an association node, wherein the graph database comprises the stored core node of the existing mesh point and the stored attribute association node of the existing mesh point; the attribute corresponds to a signaling data type; and the ith core node and the m associated nodes have a mapping relation, wherein i is more than or equal to 1 and less than or equal to n, i is an integer, n is the total number of nodes in the site selection information association diagram, and m is the number of the types of the signaling data.
According to the embodiment of the present disclosure, the graph database further includes stored service level data of existing websites, and the preprocessing of the local website addressing information association map includes: normalizing the signaling data of the ith core node to obtain a feature vector of the ith core node, wherein the feature vector of the ith core node comprises m features, and the m features are obtained based on an associated node which has a mapping relation with the ith core node; acquiring an edge feature vector based on the distance between the ith core node and the rest n-1 core nodes, and acquiring an adjacent matrix of the ith core node based on the edge feature vector; marking the core node data of the existing network node as service grade data; and acquiring local site selection information association diagram data based on the feature vector of the core node, the adjacency matrix of the core node and the existing site core node data.
According to an embodiment of the present disclosure, the obtaining of the prediction result of site selection of a mesh point further includes: acquiring service grade prediction grades of k target areas, wherein k is an integer larger than 1; and sequencing the service grade prediction grades of the k target areas from high to low, and taking the first sequenced target area as an optimal mesh point prediction site.
According to the embodiment of the disclosure, the service data includes at least two of service transaction data, customer growth data and customer flow data, and the service data is used for acquiring service level data of the existing network point.
According to an embodiment of the present disclosure, the site selection prediction model includes a first graph convolution network layer, a second graph convolution network layer, a graph feature averaging layer, a first full-link layer, a second full-link layer, and a classification network layer.
According to an embodiment of the present disclosure, the obtaining of the prediction result of site selection of a mesh point includes: inputting the local network point addressing information association graph data to a first graph convolution network layer to obtain first-order node association characteristic information; inputting the first-order node correlation characteristic information into a second graph convolution network layer to obtain second-order node correlation characteristic information; inputting the second-order node correlation characteristic information to a graph characteristic average layer to obtain graph average characteristic information; inputting the graph average feature information to a first full-connection layer to obtain first global feature information; inputting the first global feature information to a second full-connection layer after nonlinear activation transformation to obtain second global feature information; and inputting the second global feature information into a classification network layer to obtain a target area service level prediction result.
According to the embodiment of the disclosure, when the site selection prediction model is pre-trained, parameters of the site selection prediction model to be trained are updated by adopting a back propagation and gradient descent optimization algorithm until a preset training cutoff condition is reached.
A second aspect of the present disclosure provides a mesh point addressing prediction apparatus, including: the first acquisition module is configured to acquire signaling data of a target area; the first calculation module is configured to store the signaling data of the target area into a graph database and construct a site selection information association graph; the second calculation module is configured to divide the network point addressing information association graph based on preset distance granularity and acquire m pieces of local network point addressing information association graph data, wherein m is an integer greater than or equal to 2; and the prediction module is configured to input the preprocessed local site selection information association graph into a pre-trained site selection prediction model to obtain a site selection prediction result, wherein the site selection prediction result comprises a service grade prediction grade of a target area, and the pre-trained site selection prediction model is constructed on the basis of a graph convolution neural network.
According to an embodiment of the present disclosure, the first calculation module further includes a storage submodule configured to store the target area and the attributes of the target area into a graph database with the target area as a core node and the attributes of the target area as associated nodes. The graph database comprises stored core nodes of the existing mesh points and attribute association nodes of the existing mesh points; the attribute corresponds to a signaling data type; the ith core node and the m associated nodes have a mapping relation, wherein i satisfies that i is more than or equal to 1 and less than or equal to n, i is an integer, n is the total number of nodes in the site selection information associated graph, and m is the number of types of signaling data.
According to an embodiment of the disclosure, the prediction module includes a first convolution sub-module, a second convolution sub-module, a feature averaging sub-module, a first synthesis sub-module, a second synthesis sub-module, and a classification sub-module. The first convolution submodule is configured to input all m local network point addressing information correlation diagram data to a first diagram convolution network layer, and obtain first-order node correlation characteristic information. And the second convolution sub-module is configured to input the first-order node correlation characteristic information to a second graph convolution network layer to acquire second-order node correlation characteristic information. And the characteristic averaging submodule is configured to input the second-order node correlation characteristic information to a graph characteristic averaging layer to acquire graph average characteristic information. The first synthesis submodule is configured to input the graph average feature information to a first full-connection layer, and obtain first global feature information. And the second synthesis submodule is configured to input the first global feature information to a second full-connection layer after nonlinear activation transformation, and obtain second global feature information. And the classification sub-module is configured to input the second global feature information into a classification network layer to obtain a network point service level prediction result.
According to the embodiment of the disclosure, the graph database further comprises stored service level data of existing network points, and the prediction module further comprises a normalization sub-module, a vectorization sub-module, a marking sub-module and a data summarization sub-module. The normalization submodule is configured to perform normalization processing on signaling data of an ith core node to obtain a feature vector of the ith core node, wherein the feature vector of the ith core node includes m features, and the m features are obtained based on an associated node having a mapping relationship with the ith core node. The vectorization sub-module is configured to obtain an edge feature vector based on the distance between the ith core node and the rest n-1 core nodes, and obtain an adjacency matrix of the ith core node based on the edge feature vector. The marking submodule is configured to mark the existing mesh point core node data as service level data. The data summarization submodule is configured to acquire local site addressing information association diagram data based on the feature vector of the core node, the adjacency matrix of the core node and the existing site core node data.
According to the embodiment of the disclosure, the mesh point site selection prediction device further comprises a sorting module. The ranking module is configured to obtain traffic class prediction ratings for k target regions, k being an integer greater than 1. And sequencing the service grade prediction grades of the k target areas from high to low, and taking the first sequenced target area as an optimal mesh point prediction site.
A third aspect of the present disclosure provides an electronic device, comprising: one or more processors; a memory for storing one or more programs, wherein the one or more programs, when executed by the one or more processors, cause the one or more processors to perform the mesh point siting prediction method described above.
A fourth aspect of the present disclosure also provides a computer-readable storage medium having stored thereon executable instructions that, when executed by a processor, cause the processor to perform the above-described mesh point addressing prediction method.
A fifth aspect of the present disclosure also provides a computer program product comprising a computer program which, when executed by a processor, implements the above-described mesh point addressing prediction method.
According to the method provided by the embodiment of the disclosure, the site selection prediction model is built through the neural network of the graph, the incidence relation among the characteristic data is fully mined, the problems of low site selection efficiency and low accuracy of the traditional bank site are solved, and a more scientific and accurate site selection prediction method is provided.
Drawings
The foregoing and other objects, features and advantages of the disclosure will be apparent from the following description of embodiments of the disclosure, which proceeds with reference to the accompanying drawings, in which:
fig. 1 schematically illustrates an application scenario diagram of a mesh point addressing prediction method, apparatus, device, medium, and program product according to an embodiment of the disclosure.
Fig. 2 schematically shows a flowchart of a mesh point addressing prediction method according to an embodiment of the present disclosure.
Fig. 3 schematically shows a flowchart of a method for constructing a website addressing information association map according to an embodiment of the present disclosure.
Figure 4 schematically illustrates a flow chart of a method of pre-processing a local site addressing information correlation map in accordance with an embodiment of the present disclosure.
Fig. 5 schematically shows a flowchart of a method for obtaining a website addressing prediction result according to an embodiment of the present disclosure.
Fig. 6 schematically shows a flowchart of a method for obtaining a mesh point addressing prediction result based on a mesh point addressing prediction model according to an embodiment of the present disclosure.
Fig. 7 schematically shows a block diagram of a mesh point addressing prediction apparatus according to an embodiment of the present disclosure.
Fig. 8 schematically shows a block diagram of the first computing module 720 according to an embodiment of the present disclosure.
Fig. 9 schematically shows a block diagram of a prediction module 740 according to an embodiment of the present disclosure.
FIG. 10 schematically illustrates a block diagram of further prediction modules 740, according to an embodiment of the disclosure.
Fig. 11 is a block diagram schematically illustrating a mesh point addressing prediction apparatus 700 according to further embodiments of the present disclosure.
Fig. 12 schematically illustrates a block diagram of an electronic device suitable for implementing a mesh point addressing prediction method according to an embodiment of the present disclosure.
Detailed Description
Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. It should be understood that the description is illustrative only and is not intended to limit the scope of the present disclosure. In the following detailed description, for purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of the embodiments of the disclosure. It may be evident, however, that one or more embodiments may be practiced without these specific details. Moreover, in the following description, descriptions of well-known structures and techniques are omitted so as to not unnecessarily obscure the concepts of the present disclosure.
The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. The terms "comprises," "comprising," and the like, as used herein, specify the presence of stated features, steps, operations, and/or components, but do not preclude the presence or addition of one or more other features, steps, operations, or components.
All terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art unless otherwise defined. It is noted that the terms used herein should be interpreted as having a meaning that is consistent with the context of this specification and should not be interpreted in an idealized or overly formal sense.
Where a convention analogous to "at least one of A, B and C, etc." is used, in general such a construction is intended in the sense one having skill in the art would understand the convention (e.g., "a system having at least one of A, B and C" would include but not be limited to systems that have a alone, B alone, C alone, a and B together, a and C together, B and C together, and/or A, B, C together, etc.).
In the technical scheme of the disclosure, the collection, storage, use, processing, transmission, provision, disclosure, application and other processing of the personal information of the related user are all in accordance with the regulations of related laws and regulations, necessary confidentiality measures are taken, and the customs of the public order is not violated.
In the technical scheme of the disclosure, before the personal information of the user is acquired or collected, the authorization or the consent of the user is acquired.
With the continuous development of social economy, the competition of banks is increasingly intense. The traditional physical bank outlets are one of important marketing channels and the most visual way for displaying the brand images of banks, and the address selection of the physical bank outlets is directly related to the development and economic benefits of the banks. The selection of a scientific and reasonable website address becomes an important condition for the banks to gain advantages in the financial market competition, and has important significance for the development of the banks and the improvement of the core competitiveness. Therefore, establishing a rapid, accurate and automatic site selection evaluation and recommendation method for banking outlets is very important for seizing regional financial markets and competing for customer resources. The existing address selection method of bank outlets mainly comprises two methods, namely a first manual address selection method and a second method for constructing an address selection model by using signaling data according to a specified weight formula or by establishing a logistic regression machine learning mode. The first manual site selection method is characterized in that data such as people flow and traffic are collected manually, then an expert establishes a site selection model through the data to select a site, the time is long, the error of a survey result is large, the site selection is time-consuming, labor-consuming and low in efficiency, a plurality of unreasonable places can be provided, the resource repetition and waste can be caused, and the bank income is poor. Secondly, a site selection model is established by using signaling data according to an assigned weight formula or by establishing a logistic regression machine learning mode, so that the relation between the signaling data and the banking business value can be mined, the problem of high labor cost of traditional site selection is solved to a certain extent, but the assigned weight formula is often set through manual experience, and the influence of subjective judgment can cause the site selection result to be inaccurate. By establishing a simple model for logistic regression machine learning, the correlation information between feature data cannot be fully mined, so that the accuracy of site selection prediction of network points can be influenced.
Graph structure data is a more complex data structure than linear tables and trees. In a linear table, data elements are concatenated with only a linear relationship, each data element having only one direct predecessor and one direct successor. In the tree structure, there is a distinct hierarchical relationship between data elements, and a data element on each level may be related to multiple elements in the next level, but only to one element in the previous level. In the graph structure, the relationship between nodes may be arbitrary, and any two data elements in the graph may be related to each other. A graph convolution neural network is a variant of a convolution neural network on graph data that learns a representation of the graph data by stacking several first-order spectral filters before a non-linear function. The graph convolution neural network can naturally fuse attribute information of the graph to learn and mine the association between the nodes of the graph. In the process of site selection of bank network points, a large amount of unstructured signaling data needs to be collected, and the unstructured signaling data have the characteristics of unstructured, disordered and random.
The embodiment of the disclosure provides a network point site selection prediction method, which comprises the steps of obtaining signaling data of a target area; storing the signaling data of the target area into a graph database to construct a site selection information association graph; dividing the network point addressing information association diagram based on preset distance granularity to obtain m local network point addressing information association diagram data, wherein m is an integer greater than or equal to 2; and preprocessing the local site selection information association graph and inputting the preprocessed local site selection information association graph into a pre-trained site selection prediction model to obtain a site selection prediction result, wherein the site selection prediction result comprises a service grade prediction grade of a target area, and the pre-trained site selection prediction model is constructed on the basis of a graph convolution neural network.
It should be noted that the website selection prediction method, apparatus, device, medium, and program product provided in the embodiments of the present disclosure may be used in the relevant aspects of the artificial intelligence technology in the application of the graph neural network, and may also be used in various fields other than the artificial intelligence technology, such as the financial field. The application fields of the method, the device, the equipment, the medium and the program product for predicting the site selection of the network points provided by the embodiment of the disclosure are not limited.
The above-described operations for carrying out at least one of the objects of the present disclosure will be described with reference to the accompanying drawings and description thereof.
Fig. 1 schematically illustrates an application scenario diagram of a mesh point addressing prediction method, apparatus, device, medium, and program product according to an embodiment of the disclosure.
As shown in fig. 1, the application scenario 100 according to this embodiment may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 serves as a medium for providing communication links between the terminal devices 101, 102, 103 and the server 105. Network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, to name a few.
The user may use the terminal devices 101, 102, 103 to interact with the server 105 via the network 104 to receive or send messages or the like. The terminal devices 101, 102, 103 may have installed thereon various communication client applications, such as shopping-like applications, web browser applications, search-like applications, instant messaging tools, mailbox clients, social platform software, etc. (by way of example only).
The terminal devices 101, 102, 103 may be various electronic devices having display screens and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, and the like, and the terminal devices 101, 102, 103 may send signaling data to the server 105 and receive site selection prediction results.
The server 105 may be a server providing various services, such as a background management server (for example only) providing support for websites browsed by users using the terminal devices 101, 102, 103. The background management server may analyze and perform other processing on the received data such as the user request, and feed back a processing result (e.g., a webpage, information, or data obtained or generated according to the user request) to the terminal device.
It should be noted that the website selection prediction method provided by the embodiment of the present disclosure may be generally executed by the server 105. Accordingly, the website selection prediction apparatus provided by the embodiment of the present disclosure may be generally disposed in the server 105. The website selection prediction method provided by the embodiment of the present disclosure may also be executed by a server or a server cluster that is different from the server 105 and is capable of communicating with the terminal devices 101, 102, 103 and/or the server 105. Accordingly, the website selection prediction apparatus provided in the embodiment of the present disclosure may also be disposed in a server or a server cluster different from the server 105 and capable of communicating with the terminal devices 101, 102, 103 and/or the server 105.
It should be understood that the number of terminal devices, networks, and servers in fig. 1 is merely illustrative. There may be any number of terminal devices, networks, and servers, as desired for implementation.
The mesh point addressing prediction method of the disclosed embodiment will be described in detail below with reference to fig. 2 to 6 based on the scenario described in fig. 1.
Fig. 2 schematically shows a flowchart of a mesh point addressing prediction method according to an embodiment of the present disclosure.
As shown in fig. 2, the website selection prediction method of this embodiment includes operations S210 to S240, and the transaction processing method may be executed by a processor, or may be executed by any electronic device including a processor.
In operation S210, signaling data of a target area is acquired.
According to an embodiment of the present disclosure, the target area may be a pre-selected candidate area that may be a site of a website. The target area may be pre-selected based on prior investigations or expert experience. The manner in which the target area is selected is outside the scope of the discussion of the embodiments of the present disclosure. The signaling data is data related to the service level of a network point, and can be acquired by communication companies, government departments, cooperative data companies and other units, or acquired by early-stage visiting and researching. It should be noted that, in the embodiment of the present disclosure, when the signaling data includes the user information, before obtaining the user information, the user may obtain the consent or authorization. For example, a request for obtaining user information may be issued to the user before operation S210. In case that the user information can be acquired with the user' S consent or authority, the operation S210 is performed.
In operation S220, the signaling data of the target area is stored in a graph database to construct a website addressing information association graph.
According to an embodiment of the present disclosure, collected signaling data may be stored to a graph database to construct a site selection information association graph. A typical graph database may include Neo 4J. The storage process of the graph may include: reading signaling data, using the node type corresponding to the field name and the node corresponding to the field value, establishing association by using a target area or an existing network node as a core node and using node attributes (such as signaling data) as slave nodes, and storing the association in a graph database. When new node data is stored, whether the node exists or not is inquired, if the node exists, the data is not repeatedly stored in the graph database, and the target area and all the existing mesh point data are stored in the graph data according to the processing process for subsequent utilization. It can be understood that, when the embodiments of the present disclosure are used to evaluate a target area to obtain a site selection prediction result, the database includes stored attribute data of an existing site, where the attribute data of the existing site includes signaling data of the existing site.
In operation S230, the mesh point addressing information association map is segmented based on the preset distance granularity, and m pieces of local mesh point addressing information association map data are obtained, where m is an integer greater than or equal to 2.
According to the embodiment of the disclosure, in order to fully mine the association relationship between core nodes and the influence of node characteristics on the node service level, the site selection information association diagram can be segmented based on a preset distance granularity. The preset distance granularity can be set according to expert experience, for example, the site selection information association graph can be segmented by taking a core node as a center and taking 5 kilometers as a radius to obtain a local site selection information association graph, so that the influence of other sites and site attribute data within 5 kilometers on the service level of a site to be evaluated is fully mined.
In operation S240, the local site selection information association map is preprocessed and then input to a pre-trained site selection prediction model, so as to obtain a site selection prediction result. The site selection prediction result of the network points comprises the service grade prediction rating of the target area, and the pre-trained network point site selection prediction model is constructed based on a graph convolution neural network.
According to the embodiment of the disclosure, in order to facilitate the website selection prediction model to read and process data, the original unstructured signaling data can be preprocessed to be converted into a form that can be input by the model. In the processing of graph data of the present disclosure, node vector conversion and acquisition of an adjacency matrix may be performed. Wherein the vector conversion is used for vectorizing the node features, and the adjacency matrix is used for representing the association between the edges formed by the core nodes. The site selection prediction model of the net points is constructed based on a graph convolution neural network. It should be understood that when the preprocessed local network point locating information association diagram is input to the network point locating prediction model, the existing network point in the core node also has associated signaling data, and the existing network point also includes a service level label, so that a target area service prediction rating can be obtained through prediction after model processing to obtain a network point locating prediction result. In the embodiment of the disclosure, unstructured, unordered and random target area signaling data are stored as a graph data structure, and a graph convolution network is utilized to break through the dependency relationship of a two-dimensional structure, so that the associated information between feature data is more abundantly mined. The site selection prediction of the network points is converted into a classification problem, and the prediction rating of the target area service is obtained, so that the site selection deployment of the network points can be more scientific and accurate compared with the traditional method.
Embodiments of the present disclosure find that the signaling data may include at least two of economic data, geographic data, physical condition data, and peer competition data based on considerations of data accessibility, acquirability, mesh point traffic level prediction effectiveness, and data processing efficiency. Wherein the economic data comprises signaling data related to the economic development level of the target area or existing mesh point. The geographic data contains signaling data related to the geographic location of the target area or existing mesh point. The physical condition data comprises the physical attribute of the target area or the existing network, the signaling data related to the infrastructure condition and the like, and the peer competition data comprises the data of other bank networks having peer competition relationship with the target area or the existing network. Preferably, the signaling data can include a total amount of surrounding economy B, a pedestrian volume V, an age composition Y, a per-capitalized income I, a number of surrounding permanent population R, a daily takeout number T of permanent population per person, a number of surrounding elevator cells F, a number of surrounding non-elevator cells N, a number of pre-door parking spaces P, public traffic data and subway data L within 300 meters, a geographic location G of a target area or a website, a newness degree O of the target area or the website, an outdoor advertising effect K, a target area or website annual rent M, a target area or a website area C, a target area or a website internal facility D, a target area or website employee salary sum E, a number of surrounding same industry websites H, and the like. Wherein the perimeter may be defined based on a predetermined geographic range, such as a specified number of blocks; or the mesh point or the target area is taken as the center of a circle, the preset distance is taken as the radius to define the geographical range, and the like.
Fig. 3 schematically shows a flowchart of a method for constructing a website addressing information association map according to an embodiment of the present disclosure.
As shown in fig. 3, the method for constructing a website addressing information association map of this embodiment includes operation S310.
In operation S310, a target area is used as a core node, an attribute of the target area is used as a correlation node, and the target area and the attribute of the target area are stored in a graph database. Wherein the graph database comprises stored existing mesh point core nodes and existing mesh point attribute association nodes. The attributes correspond to the signaling data categories. For example, each type of signaling data is an attribute. And the ith core node and the m associated nodes have a mapping relation. Wherein i is more than or equal to 1 and less than or equal to n, i is an integer, n is the total number of nodes in the site selection information association graph, and m is the number of signaling data types. Thus, each core node forms a one-to-many association with signaling data associated therewith. In the embodiment of the disclosure, i can traverse the values of 1 to n to obtain a website addressing information association graph for expressing the association relationship of all nodes.
According to embodiments of the present disclosure, the graph database also includes stored traffic class data for existing mesh points. When the site selection prediction model of the embodiment of the disclosure is applied, the prediction of the service level of the target area can be realized by mining the incidence relation between the service level data of the existing site and the signaling data of the existing site.
Fig. 4 schematically shows a flowchart of a method for preprocessing a local site addressing information association graph according to an embodiment of the present disclosure.
As shown in fig. 4, the method for preprocessing the local site addressing information correlation map of this embodiment includes operations S410 to S440.
In operation S410, the signaling data of the ith core node is normalized to obtain a feature vector of the ith core node. The ith core node feature vector comprises m features, and the m features are obtained based on the associated nodes which have a mapping relation with the ith core node.
In the embodiment of the disclosure, to avoid the problem of difficult convergence caused by too large difference between the values of the signaling data, the preprocessing method may include a step of normalizing the signaling data. In particular, the signaling data may be normalized using Z-score normalization. This is not true of formula (1):
Figure BDA0003602559430000131
wherein X is the value of the signalling data X, muxIs the mean, σ, of the signalling data XxIs the standard deviation of the signaling data X. It should be appreciated that before normalizing the signaling data, feature engineering may also be utilized to perform data washing on the signaling data, including mean filling and outlier sample deletion. The geographic location G of the target area or the website can be represented based on latitude and longitude, the recency degree O of the target area or the website can be represented by value assignment, the outdoor advertising effect K can be represented by value assignment, and the internal facilities D of the target area or the website can be represented by counting the number of internal facility types or presetting the number of internal facility types to realize the feature representation.
In operation S420, an edge feature vector is obtained based on distances between the ith core node and the remaining n-1 core nodes, and an adjacency matrix of the ith core node is obtained based on the edge feature vector.
In the embodiment of the disclosure, in order to mine the association relationship between the target area and other mesh points which may have an influence on the service level of the target area due to the geographic position association with the target area, edge feature vectors may be obtained according to the distance between the ith core node and the remaining n-1 core nodes, and an adjacency matrix is further constructed. Illustratively, the mesh point addressing information association map of the embodiment of the present disclosure is undirected graph G ═ V, E. In undirected graph G, there is a node vxE.g. V, edge (V)i，vj) E, where i, j ∈ x. The adjacency matrix of the core node may be as shown in equation (2):
Figure BDA0003602559430000141
wherein, Aij represents the values of two nodes i and j in the adjacency matrix, vi represents the node i, vj represents the node j, and E represents the edge set. Therefore, the influence of the distance between the nodes on the service level of the nodes can be mined.
In operation S430, existing mesh point core node data is marked as service level data. According to an embodiment of the present disclosure, the traffic class data of the existing mesh points in the core node may be marked as a core node value. Therefore, after normalization, the distance between a certain core node and another core node is used as the feature vector [ S ] of each edge in the graph data, the feature vector [ S ] is converted into an adjacent matrix of the core node, and each existing core node data in each local graph is marked with the service level data of the node.
In operation S440, local mesh point addressing information association map data is acquired based on the feature vector of the core node, the adjacency matrix of the core node, and the existing mesh point core node data. It can be understood that the local mesh point addressing information association graph data includes feature vectors of core nodes, adjacency matrices of core nodes, and service level data of the core nodes of the existing mesh points. After the local network point addressing information association diagram data is processed by a network point prediction model, a target area service level prediction rating result can be output.
In some embodiments, the business data comprises at least two of business transaction data, customer growth data, and customer traffic data. The service level of the existing network point can be evaluated based on the service data, and the service level data of the existing network point can be obtained. For example, the mesh point traffic level may be scored in percent, may be graded by 5 into one grade, and may be graded by 20. The grade corresponding score can be flexibly set based on expert experience and combined with service requirements.
According to the embodiment of the disclosure, the candidate target areas may include a plurality of target areas, and accordingly, after the service level prediction rating results of the plurality of target areas are obtained, further screening may be performed to obtain an optimal site selection site for mesh point prediction.
Fig. 5 schematically shows a flowchart of a method for obtaining a mesh point addressing prediction result according to an embodiment of the present disclosure.
As shown in fig. 5, the method for obtaining the site selection prediction result of this embodiment includes operations S510 to S520.
In operation S510, service level prediction ratings of k target areas are obtained, where k is an integer greater than 1.
In operation S520, the service level prediction ratings of the k target areas are ranked in order from high to low, and the first target area in the ranking is used as an optimal mesh point prediction site. It should be understood that k is the total number of candidate target regions.
In the embodiment of the disclosure, the network point addressing prediction model is constructed based on a graph convolution neural network and comprises a first graph convolution network layer, a second graph convolution network layer, a graph characteristic average layer, a first full-connection layer, a second full-connection layer and a classification network layer.
Fig. 6 schematically shows a flowchart of a method for obtaining a mesh point addressing prediction result based on a mesh point addressing prediction model according to an embodiment of the present disclosure.
As shown in fig. 6, the method for obtaining a mesh point addressing prediction result based on the mesh point addressing prediction model of this embodiment includes operations S610 to S660.
In operation S610, the local site addressing information association graph data is input to the first graph convolution network layer, and first-order node association feature information is obtained.
In operation S620, the first-order node-associated feature information is input to the second graph convolution network layer, and second-order node-associated feature information is obtained.
In operation S630, the second-order node-associated feature information is input to a graph feature averaging layer, and graph average feature information is obtained.
In operation S640, the graph average feature information is input to the first fully connected layer, and first global feature information is obtained.
In operation S650, the first global feature information is input to the second fully-connected layer after being subjected to nonlinear activation transformation, and second global feature information is obtained.
In operation S660, the second global feature information is input to the classification network layer, and a target area service level prediction result is obtained.
In the embodiment of the present disclosure, the local mesh point addressing information association map data is a local mesh point addressing information association map data test set including a target area. The model firstly extracts first-order and second-order node association characteristic information through two graph convolution network layers. Because the number of nodes in different local site addressing information association graphs is different, the second-order node association characteristic information can be averaged based on a graph characteristic averaging layer. Further, the averaged image average feature information is integrated with global features by adopting two fully-connected layers. After the first global feature information is obtained through the first layer full connection layer, the first global feature information may be subjected to nonlinear activation transformation by using a ReLU function. And inputting second global feature information obtained after the processing of the two fully-connected layers into a classification network layer, and calculating the probability of predicting each service level of the target area after linear transformation is carried out on the second global feature information by using a classifier of the classification layer to obtain a classification result, namely a service level prediction result of the target area.
It should be understood that, before the model is applied, a part of data only including existing site signaling data and service classes may be used to construct and locally divide a sample site addressing information association graph, and a training graph data set is constructed after preprocessing a divided local graph data sample, and another part of sample data only including existing site signaling data and service class data is input as a verification graph data set for pre-training the site addressing prediction model to be trained. The construction, segmentation and preprocessing method of the graph data are described in the foregoing, and are not described herein again. The ratio of the training diagram data and the verification diagram data may be preset based on expert experience.
According to the embodiment of the disclosure, when the site selection prediction model is pre-trained, parameters of the site selection prediction model to be trained are updated by adopting a back propagation and gradient descent optimization algorithm until a preset training cutoff condition is reached. In some specific embodiments, a Nadam gradient descent optimization algorithm may be employed to improve data processing efficiency.
In some specific embodiments, the training graph data set may be input to the model to be trained in batches while training the model. An example data sheet exchange process is as follows: for undirected graph G ═ (V, E), there is each node VxE.g. V, each edge (V)i，vj) E, where i, j E x. Defining the hidden state of each node as hvxThe neighbor set of each node is N (v)x) For each node v in the graphxThe information propagation formula in the model is shown as formula (3):
Figure BDA0003602559430000161
wherein c is the number of neural network layers of the graph
Figure BDA0003602559430000162
For the feature of the neighboring node u at the c level, | N (v)x) I is the number of neighbor nodes, w(c)Is a weight vector on level c, b(c)ReLU is the activation function for the bias term on layer c.
What is obtained by the graph neural network layer is the characteristics of each core node, and the ultimate task is to classify the graph data. Since the number of core nodes of each local graph may be different in the embodiment of the present disclosure, the graph features read out after graph convolution are averaged by using the graph feature averaging layer to obtain an average graph feature hgAs shown in formula (4):
Figure BDA0003602559430000171
wherein h isuIs the graph feature vector of the neighboring node u, | n (v) | is the number of nodes.
And (3) converting the average graph characteristics through a full connection layer and then inputting the average graph characteristics into a classification network layer for service grade classification, wherein the formula (5) is as follows:
Figure BDA0003602559430000172
wherein, yiTo predict the probability of belonging to a class (i.e., traffic class) i, M is the number of classes, wiIs a weight vector, biAs an offset term, hgIs an average graph feature.
In embodiments of the present disclosure, the classification network layer may include a Softmax layer. A cross entropy loss function may be employed in the training. The training cutoff condition may be setting a preset training round, or may be reaching a preset model precision. In the training process, the number of layers of a graph neural network layer and a full connection layer is adjusted, and the distance granularity adjustment comparison of local site selection information association graph segmentation is found, 5Km is adopted as the distance granularity for segmentation, a model adopts a 2-layer graph neural network layer and a 2-layer full connection layer, and ReLU transformation is used after the 1 st full connection layer to obtain higher model accuracy, so that the scientificity and accuracy of site selection prediction are improved.
Based on the site selection prediction method, the disclosure also provides a site selection prediction device. The apparatus will be described in detail below with reference to fig. 7 to 11.
Fig. 7 schematically shows a block diagram of a mesh point addressing prediction apparatus according to an embodiment of the present disclosure.
As shown in fig. 7, the mesh point addressing prediction apparatus 700 of this embodiment includes a first obtaining module 710, a first calculating module 720, a second calculating module 730, and a prediction module 740.
Wherein the first obtaining module 710 is configured to obtain the signaling data of the target area.
The first calculation module 720 is configured to store the signaling data of the target area into a graph database to construct a site selection information association graph.
The second calculating module 730 is configured to segment the mesh point addressing information association map based on a preset distance granularity, and obtain m pieces of local mesh point addressing information association map data, where m is an integer greater than or equal to 2.
The prediction module 740 is configured to input the preprocessed local site selection information association diagram to a pre-trained site selection prediction model to obtain a site selection prediction result. The site selection prediction result of the network points comprises the service grade prediction rating of the target area, and the pre-trained network point site selection prediction model is constructed based on a graph convolution neural network.
In some particular embodiments, the first computing module 720 may also include a storage submodule 7201.
Fig. 8 schematically shows a block diagram of the first computing module 720 according to an embodiment of the present disclosure.
As shown in fig. 8, the storage submodule 7201 is configured to store the target area and the attribute of the target area in the graph database, with the target area as a core node and the attribute of the target area as an association node. The graph database comprises stored existing mesh point core nodes and existing mesh point attribute association nodes; the attribute corresponds to a signaling data type; and the ith core node and the m associated nodes have a mapping relation, wherein i is more than or equal to 1 and less than or equal to n, i is an integer, n is the total number of nodes in the site selection information association diagram, and m is the number of the types of the signaling data.
In some specific embodiments, the prediction module 740 may further include a first convolution sub-module 7401, a second convolution sub-module 7402, a feature averaging sub-module 7403, a first synthesis sub-module 7404, a second synthesis sub-module 7405, and a classification sub-module 7406.
Fig. 9 schematically shows a block diagram of a prediction module 740 according to an embodiment of the present disclosure.
As shown in fig. 9, the first convolution sub-module 7401 is configured to input all m local mesh point addressing information correlation map data into the first map convolution network layer, and obtain first-order node correlation characteristic information.
The second convolution submodule 7402 is configured to input the first order node correlation characteristic information to the second graph convolution network layer, and obtain second order node correlation characteristic information.
The feature averaging submodule 7403 is configured to input the second-order node-associated feature information to the graph feature averaging layer, and obtain graph average feature information.
The first synthesis sub-module 7404 is configured to input the graph mean feature information to the first fully connected layer, obtaining first global feature information.
The second integrating sub-module 7405 is configured to input the first global feature information to the second fully connected layer after the first global feature information is subjected to nonlinear activation transformation, so as to obtain second global feature information.
The classification submodule 7406 is configured to input the second global feature information to the classification network layer, and obtain a prediction result of the website service level.
In other specific embodiments, the graph database further includes stored traffic class data for existing mesh points, and the prediction module 740 further includes a normalization sub-module 7407, a vectorization sub-module 7408, a labeling sub-module 7409, and a data summarization sub-module 7410.
FIG. 10 schematically illustrates a block diagram of further prediction modules 740, according to an embodiment of the disclosure.
As shown in fig. 10, the normalization submodule 7407 is configured to normalize the signaling data of the ith core node, and obtain a feature vector of the ith core node, where the feature vector of the ith core node includes m features, and the m features are obtained based on an associated node having a mapping relationship with the ith core node.
The vectorization submodule 7408 is configured to obtain an edge feature vector based on the distance of the ith core node from the remaining n-1 core nodes, and obtain an adjacency matrix for the ith core node based on the edge feature vector.
The marking submodule 7409 is configured to mark existing mesh point core node data as traffic class data.
The data summarization sub-module 7410 is configured to obtain local mesh point addressing information association map data based on the feature vectors of the core nodes, the adjacency matrix of the core nodes, and the existing mesh point core node data.
According to other embodiments of the present disclosure, the mesh point addressing prediction apparatus further comprises a ranking module 750.
Fig. 11 schematically illustrates a block diagram of a mesh point addressing prediction apparatus 700 according to further embodiments of the present disclosure.
As shown in fig. 11, the ranking module 750 is configured to obtain the traffic class prediction ratings of k target areas, k being an integer greater than 1. And sequencing the service grade prediction grades of the k target areas from high to low, and taking the first sequenced target area as an optimal mesh point prediction site.
According to the embodiment of the present disclosure, any plurality of the first obtaining module 710, the first calculating module 720, the second calculating module 730, the predicting module 740, the sorting module 750, the storage sub-module 7201, the first convolution sub-module 7401, the second convolution sub-module 7402, the feature averaging sub-module 7403, the first synthesis sub-module 7404, the second synthesis sub-module 7405, the classification sub-module 7406, the normalization sub-module 7407, the vectorization sub-module 7408, the marking sub-module 7409, and the data summarization sub-module 7410 may be combined into one module to be implemented, or any one of them may be split into a plurality of modules. Alternatively, at least part of the functionality of one or more of these modules may be combined with at least part of the functionality of the other modules and implemented in one module. According to an embodiment of the disclosure, at least one of the first obtaining module 710, the first calculating module 720, the second calculating module 730, the predicting module 740, the ordering module 750, the storage sub-module 7201, the first convolution sub-module 7401, the second convolution sub-module 7402, the feature averaging sub-module 7403, the first synthesis sub-module 7404, the second synthesis sub-module 7405, the classification sub-module 7406, the normalization sub-module 7407, the vectorization sub-module 7408, the labeling sub-module 7409, and the data summarization sub-module 7410 may be implemented at least in part as a hardware circuit, such as Field Programmable Gate Arrays (FPGAs), Programmable Logic Arrays (PLAs), systems on a chip, systems on a substrate, systems on a package, Application Specific Integrated Circuits (ASICs), or may be implemented in hardware or firmware in any other reasonable way of integrating or packaging circuits, or in any one of three implementations, software, hardware and firmware, or in any suitable combination of any of them. Alternatively, at least one of the first obtaining module 710, the first calculating module 720, the second calculating module 730, the predicting module 740, the ordering module 750, the storing sub-module 7201, the first convolution sub-module 7401, the second convolution sub-module 7402, the feature averaging sub-module 7403, the first integrating sub-module 7404, the second integrating sub-module 7405, the classifying sub-module 7406, the normalizing sub-module 7407, the vectoring sub-module 7408, the marking sub-module 7409, and the data summarizing sub-module 7410 may be implemented at least partially as a computer program module, which, when executed, may perform a corresponding function.
Fig. 12 schematically illustrates a block diagram of an electronic device suitable for implementing a mesh point addressing prediction method according to an embodiment of the present disclosure.
As shown in fig. 12, an electronic apparatus 900 according to an embodiment of the present disclosure includes a processor 901 which can perform various appropriate actions and processes in accordance with a program stored in a Read Only Memory (ROM)902 or a program loaded from a storage portion 908 into a Random Access Memory (RAM) 903. Processor 901 may comprise, for example, a general purpose microprocessor (e.g., a CPU), an instruction set processor and/or associated chipset, and/or a special purpose microprocessor (e.g., an Application Specific Integrated Circuit (ASIC)), among others. The processor 901 may also include on-board memory for caching purposes. The processor 901 may comprise a single processing unit or a plurality of processing units for performing the different actions of the method flows according to embodiments of the present disclosure.
In the RAM 903, various programs and data necessary for the operation of the electronic apparatus 900 are stored. The processor 901, the ROM 902, and the RAM 903 are connected to each other through a bus 904. The processor 901 performs various operations of the method flows according to the embodiments of the present disclosure by executing programs in the ROM 902 and/or the RAM 903. Note that the programs may also be stored in one or more memories other than the ROM 902 and the RAM 903. The processor 901 may also perform various operations of the method flows according to embodiments of the present disclosure by executing programs stored in the one or more memories.
Electronic device 900 may also include input/output (I/O) interface 905, input/output (I/O) interface 905 also connected to bus 904, according to an embodiment of the present disclosure. The electronic device 900 may also include one or more of the following components connected to the I/O interface 905: an input portion 906 including a keyboard, a mouse, and the like; an output section 907 including components such as a Cathode Ray Tube (CRT), a Liquid Crystal Display (LCD), and the like, and a speaker; a storage portion 908 including a hard disk and the like; and a communication section 909 including a network interface card such as a LAN card, a modem, or the like. The communication section 909 performs communication processing via a network such as the internet. The drive 910 is also connected to the I/O interface 905 as necessary. A removable medium 911 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is mounted on the drive 910 as necessary, so that a computer program read out therefrom is mounted into the storage section 908 as necessary.
The present disclosure also provides a computer-readable storage medium, which may be contained in the apparatus/device/system described in the above embodiments; or may exist separately and not be assembled into the device/apparatus/system. The computer-readable storage medium carries one or more programs which, when executed, implement the method according to an embodiment of the disclosure.
According to embodiments of the present disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium, which may include, for example but is not limited to: a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present disclosure, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. For example, according to embodiments of the present disclosure, a computer-readable storage medium may include the ROM 902 and/or RAM 903 described above and/or one or more memories other than the ROM 902 and RAM 903.
Embodiments of the present disclosure also include a computer program product comprising a computer program containing program code for performing the method illustrated in the flow chart. When the computer program product runs in a computer system, the program code is used for causing the computer system to realize the method provided by the embodiment of the disclosure.
The computer program performs the above-described functions defined in the system/apparatus of the embodiments of the present disclosure when executed by the processor 901. The systems, apparatuses, modules, units, etc. described above may be implemented by computer program modules according to embodiments of the present disclosure.
In one embodiment, the computer program may be hosted on a tangible storage medium such as an optical storage device, a magnetic storage device, or the like. In another embodiment, the computer program may also be transmitted, distributed in the form of a signal on a network medium, and downloaded and installed through the communication section 909 and/or installed from the removable medium 911. The computer program containing program code may be transmitted using any suitable network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the foregoing.
In such an embodiment, the computer program may be downloaded and installed from a network through the communication section 909, and/or installed from the removable medium 911. The computer program, when executed by the processor 901, performs the above-described functions defined in the system of the embodiment of the present disclosure. The systems, devices, apparatuses, modules, units, etc. described above may be implemented by computer program modules according to embodiments of the present disclosure.
In accordance with embodiments of the present disclosure, program code for executing computer programs provided by embodiments of the present disclosure may be written in any combination of one or more programming languages, and in particular, these computer programs may be implemented using high level procedural and/or object oriented programming languages, and/or assembly/machine languages. The programming language includes, but is not limited to, programming languages such as Java, C + +, python, the "C" language, or the like. The program code may execute entirely on the user computing device, partly on the user device, partly on a remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computing device (e.g., through the internet using an internet service provider).
The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams or flowchart illustration, and combinations of blocks in the block diagrams or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
Those skilled in the art will appreciate that various combinations and/or combinations of features recited in the various embodiments and/or claims of the present disclosure can be made, even if such combinations or combinations are not expressly recited in the present disclosure. In particular, various combinations and/or combinations of the features recited in the various embodiments and/or claims of the present disclosure may be made without departing from the spirit or teaching of the present disclosure. All such combinations and/or associations are within the scope of the present disclosure.
The embodiments of the present disclosure have been described above. However, these examples are for illustrative purposes only and are not intended to limit the scope of the present disclosure. Although the embodiments are described separately above, this does not mean that the measures in the embodiments cannot be used in advantageous combination. The scope of the disclosure is defined by the appended claims and equivalents thereof. Various alternatives and modifications can be devised by those skilled in the art without departing from the scope of the present disclosure, and such alternatives and modifications are intended to be within the scope of the present disclosure.

Claims (13)

1. A mesh point site selection prediction method is characterized by comprising the following steps:
acquiring signaling data of a target area;
storing the signaling data of the target area into a graph database to construct a site selection information association graph;
dividing the network point addressing information association graph based on preset distance granularity to obtain m pieces of local network point addressing information association graph data, wherein m is an integer greater than or equal to 2;
and inputting the preprocessed local site selection information association diagram into a pre-trained site selection prediction model to obtain a site selection prediction result,
the site selection prediction result of the network points comprises the service grade prediction rating of the target area, and the pre-trained network point site selection prediction model is constructed based on a graph convolution neural network.
2. The method of claim 1, wherein the signaling data comprises at least two of economic data, geographic data, physical condition data, and industry competition data.
3. The method of claim 1, wherein the constructing a mesh point addressing information correlation map comprises:
storing the target area and the attributes of the target area into a graph database by taking the target area as a core node and the attributes of the target area as correlation nodes,
the graph database comprises stored existing mesh point core nodes and existing mesh point attribute association nodes; the attribute corresponds to a signaling data category; and the ith core node and the m associated nodes have a mapping relation, wherein i is more than or equal to 1 and less than or equal to n, i is an integer, n is the total number of nodes in the site selection information association diagram, and m is the number of the types of the signaling data.
4. The method of claim 3, wherein said graph database further comprises stored traffic class data of existing mesh points, said preprocessing said local mesh point addressing information association map comprising:
normalizing the signaling data of the ith core node to obtain a feature vector of the ith core node, wherein the feature vector of the ith core node comprises m features, and the m features are obtained based on associated nodes which have a mapping relation with the ith core node;
acquiring an edge feature vector based on the distance between the ith core node and the rest n-1 core nodes, and acquiring an adjacent matrix of the ith core node based on the edge feature vector;
marking the core node data of the existing network node as service grade data; and
and acquiring local mesh point addressing information association diagram data based on the feature vector of the core node, the adjacency matrix of the core node and the existing mesh point core node data.
5. The method of claim 1, wherein said obtaining a mesh point siting prediction further comprises:
acquiring service grade prediction grades of k target areas, wherein k is an integer larger than 1;
and sequencing the service grade prediction grades of the k target areas from high to low, and taking the first sequenced target area as an optimal mesh point prediction site.
6. A method according to any one of claims 4 to 5, wherein the service data comprises at least two of service transaction data, customer growth data and customer traffic data, and the service data is used for acquiring service level data of an existing website.
7. The method of claim 1, wherein the mesh point addressing prediction model comprises a first graph convolution network layer, a second graph convolution network layer, a graph feature averaging layer, a first fully connected layer, a second fully connected layer, and a classification network layer.
8. The method of claim 4, wherein the obtaining the site selection prediction result comprises:
inputting the local network point addressing information association graph data to a first graph convolution network layer to obtain first-order node association characteristic information;
inputting the first-order node correlation characteristic information into a second graph convolution network layer to obtain second-order node correlation characteristic information;
inputting the second-order node correlation characteristic information to a graph characteristic average layer to obtain graph average characteristic information;
inputting the graph average feature information to a first full-connection layer to obtain first global feature information;
inputting the first global feature information to a second full-connection layer after nonlinear activation transformation to obtain second global feature information; and
and inputting the second global feature information into a classification network layer to obtain a target area service level prediction result.
9. The method of claim 7, wherein, in pre-training the mesh point siting prediction model, parameters of the mesh point siting prediction model to be trained are updated using a back propagation and gradient descent optimization algorithm until a preset training cutoff condition is reached.
10. A mesh point addressing prediction apparatus, comprising:
the first acquisition module is configured to acquire signaling data of a target area;
the first calculation module is configured to store the signaling data of the target area into a graph database and construct a site selection information association graph;
the second calculation module is configured to segment the mesh point addressing information association map based on preset distance granularity and acquire m pieces of local mesh point addressing information association map data, wherein m is an integer greater than or equal to 2; and
and the prediction module is configured to input the preprocessed local site selection information association graph into a pre-trained site selection prediction model to obtain a site selection prediction result, wherein the site selection prediction result comprises a service grade prediction grade of a target area, and the pre-trained site selection prediction model is constructed on the basis of a graph convolution neural network.
11. An electronic device, comprising:
one or more processors;
a storage device for storing one or more programs,
wherein the one or more programs, when executed by the one or more processors, cause the one or more processors to perform the method of any of claims 1-9.
12. A computer readable storage medium having stored thereon executable instructions which, when executed by a processor, cause the processor to perform the method of any one of claims 1 to 9.
13. A computer program product comprising a computer program which, when executed by a processor, implements a method according to any one of claims 1 to 9.
CN202210408033.4A 2022-04-19 2022-04-19 Site selection prediction method, site selection prediction device, site selection prediction apparatus, site selection prediction medium, and program product Pending CN114663165A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210408033.4A CN114663165A (en) 2022-04-19 2022-04-19 Site selection prediction method, site selection prediction device, site selection prediction apparatus, site selection prediction medium, and program product

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210408033.4A CN114663165A (en) 2022-04-19 2022-04-19 Site selection prediction method, site selection prediction device, site selection prediction apparatus, site selection prediction medium, and program product

Publications (1)

Publication Number Publication Date
CN114663165A true CN114663165A (en) 2022-06-24

Family

ID=82034765

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210408033.4A Pending CN114663165A (en) 2022-04-19 2022-04-19 Site selection prediction method, site selection prediction device, site selection prediction apparatus, site selection prediction medium, and program product

Country Status (1)

Country Link
CN (1) CN114663165A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115759319A (en) * 2022-08-02 2023-03-07 日照市规划设计研究院集团有限公司 Express delivery network and transfer station site selection method considering job distribution

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115759319A (en) * 2022-08-02 2023-03-07 日照市规划设计研究院集团有限公司 Express delivery network and transfer station site selection method considering job distribution
CN115759319B (en) * 2022-08-02 2024-02-20 日照市规划设计研究院集团有限公司 Express delivery network site and transfer site selection method considering job and location distribution

Similar Documents

Publication Publication Date Title
CN111767440B (en) Vehicle portrayal method based on knowledge graph, computer equipment and storage medium
Wismans et al. Improving a priori demand estimates transport models using mobile phone data: a Rotterdam-region case
Amarsaikhan et al. Applications of remote sensing and geographic information systems for urban land-cover change studies in Mongolia
CN112418696A (en) Method and device for constructing urban traffic dynamic knowledge map
CN114416900A (en) Method and device for analyzing track stop point
CN111191825A (en) User default prediction method and device and electronic equipment
CN116542395A (en) Low-carbon building monitoring system and method
Wang et al. Measuring urban vibrancy of residential communities using big crowdsourced geotagged data
CN115605903A (en) System and method for quickly composing, launching and configuring a customizable second-level migration structure with a built-in audit and monitoring structure
Bueno et al. Application of an opinion consensus aggregation model based on OWA operators to the recommendation of tourist sites
Casali et al. A topological characterization of flooding impacts on the Zurich road network
CN114663165A (en) Site selection prediction method, site selection prediction device, site selection prediction apparatus, site selection prediction medium, and program product
Lin et al. Knowledge transfer for large‐scale urban growth modeling based on formal concept analysis
Dagan et al. What is “big data” and how should we use it? The role of large datasets, secondary data, and associated analysis techniques in outdoor recreation research
Mete et al. A Hybrid Approach for Mass Valuation of Residential Properties through Geographic Information Systems and Machine Learning Integration
Shen et al. GIS. LSP: A soft computing logic method and tool for geospatial suitability analysis
CN113254761A (en) Intelligent early warning system and method for specific behavior information
KR102103023B1 (en) System and method for providing defined contribution based unified retirement pension managment service using comparing insurance company
Zatsarinny et al. Scientific and methodological approaches to the generation of the Internet information space of scientific and educational resources
CN114493853A (en) Credit rating evaluation method, credit rating evaluation device, electronic device and storage medium
CN113191681A (en) Site selection method and device for network points, electronic equipment and readable storage medium
CN113128773A (en) Training method of address prediction model, address prediction method and device
CN113849580A (en) Subject rating prediction method and device, electronic equipment and storage medium
CN113344645A (en) House price prediction method and device and electronic equipment
CN112231299A (en) Method and device for dynamically adjusting feature library

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