CN115730147A - Customer mining method and device - Google Patents
Customer mining method and device Download PDFInfo
- Publication number
- CN115730147A CN115730147A CN202211528914.6A CN202211528914A CN115730147A CN 115730147 A CN115730147 A CN 115730147A CN 202211528914 A CN202211528914 A CN 202211528914A CN 115730147 A CN115730147 A CN 115730147A
- Authority
- CN
- China
- Prior art keywords
- node
- nodes
- neural network
- pooling
- customer
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a client mining method and a client mining device, which are applied to the technical field of artificial intelligence, wherein the method comprises the following steps: collecting client information and product information, and constructing a neural network according to the client information and the product information; extracting node characteristics of all nodes in the graph neural network through a node characteristic extraction model; calculating the similarity of the nodes according to the node characteristics, combining the nodes with the similarity reaching a preset standard into one node, and obtaining a pooling node set; and installing probes in the pooling nodes in a centralized manner, wherein the probes are used for monitoring the nodes in the pooling node in real time, acquiring the node operation in the pooling node in a centralized manner, and presetting the nodes according to the node operation and the corresponding standard. The invention can improve the analysis and prediction real-time performance, reduce the calculation dimension and improve the performance.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a client mining method and device.
Background
This section is intended to provide a background or context to the embodiments of the invention that are recited in the claims. The description herein is not admitted to be prior art by inclusion in this section.
In the field of potential customer mining and product recommendation, relevant research and application based on the graph neural network are all model analysis and prediction are carried out on fixed data within a certain time, and certain instantaneity is lacked; in the case of a huge amount of data to be studied, the computing power and performance of the system may be affected to some extent. Therefore, a real-time and efficient customer mining scheme is lacking at present.
Disclosure of Invention
The embodiment of the invention provides a client mining method, which is used for improving the efficiency of client mining and ensuring the real-time performance of client mining and comprises the following steps:
collecting client information and product information, and constructing a graph neural network according to the client information and the product information, wherein nodes in the graph neural network are clients;
extracting node characteristics of all nodes in the graph neural network through a node characteristic extraction model;
calculating the similarity of the nodes according to the node characteristics, combining the nodes with the similarity reaching a preset standard into one node, and obtaining a pooling node set;
installing probes in the pooling node set, wherein the probes are used for monitoring the nodes in the pooling node set in real time, acquiring the node states of the nodes in the pooling node set, and presetting the nodes according to the node states;
the method comprises the following steps of constructing a node feature extraction model:
selecting partial nodes in the graph neural network to form a training set;
extracting the characteristics of the nodes in the training set to obtain characteristic vectors;
reconstructing the characteristic vector through a decoder to obtain the input characteristics of the nodes in the training set;
and adjusting model parameters to enable the deviation between the input features and the feature vectors to be lower than a preset standard, so as to obtain a node feature extraction model.
The embodiment of the invention also provides a client mining device, which is used for improving the efficiency of client mining and ensuring the real-time performance of client mining, and comprises the following components:
the characteristic acquisition module is used for extracting node characteristics of all nodes in the graph neural network through the node characteristic extraction model;
the node pooling module is used for calculating the similarity of the nodes according to the node characteristics, combining the nodes with the similarity reaching a preset standard into one node and obtaining a pooled node set;
the monitoring processing module is used for installing probes in the pooling node set, the probes are used for monitoring the nodes in the pooling node set in real time, obtaining the node states of the nodes in the pooling node set, and presetting the nodes according to the node states;
the model building module is used for selecting partial nodes in the graph neural network to form a training set; extracting the characteristics of the nodes in the training set to obtain characteristic vectors; reconstructing the characteristic vector through a decoder to obtain the input characteristics of the nodes in the training set; and adjusting model parameters to enable the deviation between the input features and the feature vectors to be lower than a preset standard, so as to obtain a node feature extraction model.
The embodiment of the invention also provides computer equipment which comprises a memory, a processor and a computer program which is stored on the memory and can run on the processor, wherein the processor realizes the client mining method when executing the computer program.
An embodiment of the present invention further provides a computer-readable storage medium, where a computer program is stored, and when the computer program is executed by a processor, the client mining method is implemented.
An embodiment of the present invention further provides a computer program product, where the computer program product includes a computer program, and when the computer program is executed by a processor, the client mining method is implemented.
In the embodiment of the invention, client information and product information are collected, and a graph neural network is constructed according to the client information and the product information; extracting node characteristics of all nodes in the graph neural network through a node characteristic extraction model; calculating the similarity of the nodes according to the node characteristics, combining the nodes with the similarity reaching a preset standard into one node, and obtaining a pooling node set; installing probes in the pooling node set, wherein the probes are used for monitoring the nodes in the pooling node set in real time, acquiring node operations in the pooling node set, and presetting the nodes according to the node operations and corresponding standards; the method comprises the following steps of constructing a node feature extraction model: selecting partial nodes in the graph neural network to form a training set; extracting the characteristics of the nodes in the training set to obtain characteristic vectors; reconstructing the characteristic vector through a decoder to obtain the input characteristics of the nodes in the training set; adjusting model parameters to enable the deviation between the input features and the feature vectors to be lower than a preset standard, and obtaining a node feature extraction model; the embodiment of the invention monitors the existing customers in real time through the probe, and integrates similar data for the customers with the association relation or similar characteristics through pooling, thereby improving the analysis and prediction real-time performance, reducing the calculation dimensionality and improving the performance.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts. In the drawings:
FIG. 1 is a flow chart of a customer mining method in an embodiment of the invention;
FIG. 2 is a diagram illustrating an embodiment of pre-setting nodes according to the present invention;
FIG. 3 is an exemplary diagram of feature extraction performed on nodes in a training set according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of a customer mining device in an embodiment of the present invention;
FIG. 5 is a diagram of an embodiment of a customer mining device;
FIG. 6 is a diagram of a computer device in an embodiment of the invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the embodiments of the present invention are further described in detail below with reference to the accompanying drawings. The exemplary embodiments and descriptions of the present invention are provided to explain the present invention, but not to limit the present invention.
In the description of the present specification, the terms "comprising," "including," "having," "containing," and the like are used in an open-ended fashion, i.e., to mean including, but not limited to. Reference to the description of the terms "one embodiment," "a particular embodiment," "some embodiments," "for example," etc., means that a particular feature, structure, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. The sequence of steps involved in the various embodiments is provided to schematically illustrate the practice of the invention, and the sequence of steps is not limited and can be suitably adjusted as desired.
The following explains terms involved in the embodiments of the present invention:
pooling: in the graph neural network, nodes represent clients, and clients with similar characteristics are aggregated to form a node set.
A K-order subgraph: starting from a certain target node, a graph formed by nodes with the number of edges less than or equal to K is called a K-order subgraph of the node.
In order to solve the technical problems in the prior art, the invention provides a customer mining method which is used for monitoring the existing users, integrating similar data in a pooling manner, improving analysis and prediction timeliness and improving performance. Fig. 1 is a flowchart of a client mining method in an embodiment of the present invention, and as shown in fig. 1, the method includes the following steps:
102, extracting node characteristics of all nodes in the graph neural network through a node characteristic extraction model;
103, calculating the similarity of the nodes according to the node characteristics, combining the nodes with the similarity reaching a preset standard into one node, and obtaining a pooling node set;
and 104, installing probes in the pooling node set, wherein the probes are used for monitoring the nodes in the pooling node set in real time, acquiring the node states of the nodes in the pooling node set, and presetting the nodes according to the node states.
As can be known from the flow shown in fig. 1, the embodiment of the present invention can effectively implement customer mining; the existing customers are monitored in real time through the probes, and similar data are integrated in a pooling mode for the customers with the association relation or similar characteristics, so that the analysis and prediction real-time performance is improved, the calculation dimensionality is reduced, and the performance is improved.
In specific implementation, collecting client information and product information, and constructing a graph neural network according to the client information and the product information; extracting node characteristics of all nodes in the graph neural network through a node characteristic extraction model; and calculating the similarity of the nodes according to the node characteristics, combining the nodes with the similarity reaching a preset standard into one node, and obtaining a pooling node set.
In one embodiment, constructing a graph neural network from customer information and product information includes:
acquiring customer information and product information, and constructing a customer-product graph and a customer-customer graph;
and combining the client and product graph and the client and client graph to construct a graph neural network.
The existing data dimension reduction technology is mostly convolution processing and pooling operation, edge information is lost when convolution processes edge nodes in a graph, data filling is carried out on the edge nodes of the graph in order to acquire the edge information, the complexity is improved by the computing method, most of the pooling operation is performed according to maximum pooling and average pooling, and the similarity of the nodes is not shown.
In one embodiment, computing node similarity includes: after the node characteristics are obtained, two nodes are selected each time, the similarity of the two nodes is calculated by utilizing the cosine of the included angle, the nodes with the similarity reaching the preset standard are combined into one node to set the preset standard for reducing the output nodes in the graph neural network to the input nodes, and the times of iterative calculation of the similarity can be obtained; after the iteration is completed, similar nodes are pooled into one node.
In an embodiment, after calculating the node similarity and pooling the similar nodes, the method may further include: and installing a probe on the pooled node, monitoring the existing adjacent node of the node and the node newly associated with the node, and presetting the node according to the node operation and the corresponding standard.
Fig. 2 is an exemplary diagram illustrating a node being subjected to a preset process in the embodiment of the present invention, as shown in fig. 2, in the embodiment of the present invention, the preset process performed on the node according to the node operation and the corresponding standard may include:
Most of existing potential customer mining is based on known customer data, characteristics of known customers are analyzed, relevant information is recommended to customers related to the known customers, data instantaneity cannot be guaranteed, and in a real scene, customer node data are constantly changed and include newly added customers, inactive customers, cancelled customers and the like. In the embodiment of the invention, the existing customers are monitored in real time by using the probe so as to discover new customers, inactive customers and customers who sell accounts, and corresponding measures are taken.
In specific implementation, probes are installed in the pooling nodes in a centralized mode and used for monitoring the nodes in the pooling node in real time, obtaining node operations in the pooling node in a centralized mode, and presetting the nodes according to the node operations and corresponding standards.
In specific implementation, a probe is installed on the pooled node, and the existing adjacent node of the node and the node newly associated with the node are monitored. When a logged-off node is detected, it is deleted from the graph neural network; when detecting an inactive node, putting it into a specific region to be observed; when a new node and the node are detected to have transactions, when the number of transactions reaches a set threshold value, the new node is added into a neighbor node of the node, and the new node is marked as new.
Fig. 3 is an exemplary diagram of performing feature extraction on nodes in a training set in the embodiment of the present invention, and as shown in fig. 3, in the embodiment of the present invention, performing feature extraction on nodes in the training set to obtain a feature vector may include:
and 304, splicing the node characteristics with the node characteristics of the neighbor nodes of the target node, and processing the spliced node characteristics by using the weight matrix and the nonlinear function to obtain the characteristic vector of the target node.
During specific implementation, selecting partial nodes in the graph neural network to form a training set, selecting one node from the training set as a target node, finding out a neighbor node of the node, then recursively finding out the neighbor node of the neighbor node, and recursively finishing until the edges of the set are recurred; searching a K-order subgraph of a target node, and sequentially extracting K-1 layer characteristics and K-2 layer characteristics until the layer 1 is reached; and weighting and aggregating the characteristics of the nodes on the K-order subgraph of the target node layer by layer to form a characteristic T1, splicing the characteristic T1 with the characteristics of each node nearest to the target node, and processing by a weight matrix and a nonlinear function to obtain a characteristic vector T2 of the target node.
In specific implementation, the characteristic vector T2 is reconstructed through a decoder to obtain the input characteristics of the nodes in the training set.
In one embodiment, the method may further comprise:
when the nodes are marked as new nodes, analyzing the customer information marked as the new nodes to obtain customer characteristics;
and according to the characteristics of the client, searching and recommending the product list matched with the nodes marked as the new addition in the pooling node set.
During specific implementation, the client property, signed products and recent transaction characteristics of the newly added node client are analyzed, analyzed results are summarized, and a product list which can be matched with the newly added client in the pooled node set is found and recommended.
The embodiment of the invention also provides a client digging device, which is described in the following embodiment. Because the principle of the device for solving the problems is similar to that of a client mining method, the implementation of the device can refer to the implementation of the method, and repeated details are not repeated.
Fig. 4 is a schematic diagram of a client mining apparatus according to an embodiment of the present invention, and as shown in fig. 4, the apparatus includes:
the graph neural network building module 401 is configured to collect customer information and product information, and build a graph neural network according to the customer information and the product information, where nodes in the graph neural network are customers;
a feature acquisition module 402, configured to extract node features of all nodes in the graph neural network through a node feature extraction model;
a node pooling module 403, configured to calculate similarity of nodes according to node characteristics, and combine nodes with similarity reaching a preset standard into one node to obtain a pooled node set;
a monitoring processing module 404, configured to install a probe in the pooled node set, where the probe is used to monitor nodes in the pooled node set in real time, obtain node states of the nodes in the pooled node set, and perform preset processing on the nodes according to the node states;
the model building module 405 is used for selecting part of nodes in the neural network of the graph to form a training set; extracting the characteristics of the nodes in the training set to obtain characteristic vectors; reconstructing the characteristic vector through a decoder to obtain the input characteristics of the nodes in the training set; and adjusting model parameters to enable the deviation between the input features and the feature vectors to be lower than a preset standard, so as to obtain a node feature extraction model.
In an embodiment, the graph neural network building module 401 is specifically configured to:
acquiring customer information and product information, and constructing a customer-product graph and a customer-customer graph;
and combining the client and product graph and the client and client graph to construct a graph neural network.
In an embodiment, the node pooling module 403 is specifically configured to:
after the node characteristics are obtained, two nodes are selected each time, and the similarity of the two nodes is calculated by utilizing the cosine of the included angle.
In an embodiment, the monitoring processing module 404 is specifically configured to:
when the node state is logged off, the preset treatment is to delete the logged-off node from the graph neural network;
when the node state is inactive, the preset processing is to place the inactive node into a region to be observed of the graph neural network;
when the node state is that a transaction is carried out with a new node, the preset processing is to record the transaction times with the new node, when the transaction times reach the preset value, the new node is added into the neighbor node of the node, and the new node is marked as new.
In an embodiment, the model building module 405 is specifically configured to:
selecting a node as a target node, finding out a neighbor node of the node, and recursively finding out the neighbor node of the neighbor node until the neighbor node recurses to the edge of a set;
drawing high-order subgraphs of target nodes, and sequentially extracting the characteristics of the nodes in each layer of the high-order subgraphs;
weighting and aggregating the characteristics of the nodes in the high-order subgraph layer by layer to obtain the node characteristics of the target node;
and splicing the node characteristics with the node characteristics of the neighbor nodes of the target node, and processing the spliced node characteristics by using the weight matrix and the nonlinear function to obtain the characteristic vector of the target node.
Fig. 5 is a diagram of a specific example of a client mining device in an embodiment of the present invention, and as shown in fig. 5, in an embodiment of the present invention, the client mining device shown in fig. 4 may further include:
the product recommendation module 501 is configured to analyze the client information labeled as the newly added node when the node is labeled as the newly added node, so as to obtain client characteristics;
and according to the characteristics of the client, searching and recommending the product list matched with the nodes marked as the new addition in the pooling node set.
Fig. 6 is a schematic diagram of a computer device in an embodiment of the present invention, where the computer device 600 includes a memory 610, a processor 620, and a computer program 630 stored in the memory 610 and executable on the processor 620, and when the processor 620 executes the computer program 630, the above-mentioned client mining method is implemented.
An embodiment of the present invention further provides a computer-readable storage medium, where a computer program is stored, and when the computer program is executed by a processor, the client mining method is implemented.
An embodiment of the present invention further provides a computer program product, where the computer program product includes a computer program, and when the computer program is executed by a processor, the client mining method is implemented.
In summary, in the embodiment of the present invention, customer information and product information are collected, and a graph neural network is constructed according to the customer information and the product information; extracting node characteristics of all nodes in the graph neural network through a node characteristic extraction model; calculating the similarity of the nodes according to the node characteristics, combining the nodes with the similarity reaching a preset standard into one node, and obtaining a pooling node set; installing probes in the pooling node set, wherein the probes are used for monitoring the nodes in the pooling node set in real time, acquiring node operations in the pooling node set, and presetting the nodes according to the node operations and corresponding standards; the method comprises the following steps of constructing a node feature extraction model: selecting partial nodes in the graph neural network to form a training set; extracting the characteristics of the nodes in the training set to obtain characteristic vectors; reconstructing the characteristic vector through a decoder to obtain the input characteristics of the nodes in the training set; adjusting model parameters to enable the deviation between the input features and the feature vectors to be lower than a preset standard, and obtaining a node feature extraction model; the embodiment of the invention monitors the existing customers in real time through the probe, and integrates similar data for the customers with the association relation or similar characteristics through pooling, thereby improving the analysis and prediction real-time performance, reducing the calculation dimensionality and improving the performance.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The above-mentioned embodiments are intended to illustrate the objects, technical solutions and advantages of the present invention in further detail, and it should be understood that the above-mentioned embodiments are only exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (15)
1. A method for customer mining, comprising:
collecting client information and product information, and constructing a graph neural network according to the client information and the product information, wherein nodes in the graph neural network are clients;
extracting node characteristics of all nodes in the graph neural network through a node characteristic extraction model;
calculating the similarity of the nodes according to the node characteristics, combining the nodes with the similarity reaching a preset standard into one node, and obtaining a pooling node set;
installing probes in the pooling node set, wherein the probes are used for monitoring the nodes in the pooling node set in real time, acquiring the node states of the nodes in the pooling node set, and presetting the nodes according to the node states;
the method comprises the following steps of constructing a node feature extraction model:
selecting partial nodes in the graph neural network to form a training set;
extracting the characteristics of the nodes in the training set to obtain characteristic vectors;
reconstructing the characteristic vector through a decoder to obtain the input characteristics of the nodes in the training set;
and adjusting model parameters to enable the deviation between the input features and the feature vectors to be lower than a preset standard, so as to obtain a node feature extraction model.
2. The method of claim 1, wherein constructing a graph neural network from customer information and product information comprises:
acquiring customer information and product information, and constructing a customer-product graph and a customer-customer graph;
and combining the client and product graph and the client and client graph to construct a graph neural network.
3. The method of claim 1, wherein computing node similarities comprises:
after the node characteristics are obtained, two nodes are selected each time, and the similarity of the two nodes is calculated by utilizing the cosine of the included angle.
4. The method of claim 1, wherein pre-setting the node according to the node status comprises:
when the node state is logged off, the preset treatment is to delete the logged-off node from the graph neural network;
when the node state is inactive, the preset processing is to place the inactive node into a region to be observed of the graph neural network;
when the node state is that a transaction is carried out with a new node, the preset processing is to record the transaction times with the new node, when the transaction times reach the preset value, the new node is added into the neighbor node of the node, and the new node is marked as new.
5. The method of claim 1, wherein extracting features of nodes in the training set to obtain a feature vector comprises:
selecting a node as a target node, finding out a neighbor node of the node, and recursively finding out the neighbor node of the neighbor node until the neighbor node recurses to the edge of a set;
drawing high-order subgraphs of target nodes, and sequentially extracting the characteristics of the nodes in each layer of the high-order subgraphs;
weighting and aggregating the characteristics of the nodes in the high-order subgraph layer by layer to obtain the node characteristics of the target node;
and splicing the node characteristics with the node characteristics of the neighbor nodes of the target node, and processing the spliced node characteristics by using the weight matrix and the nonlinear function to obtain the characteristic vector of the target node.
6. The method of claim 4, further comprising:
when the nodes are marked as new, analyzing the customer information marked as the new nodes to obtain customer characteristics;
and according to the characteristics of the client, searching and recommending the product list matched with the nodes marked as the new addition in the pooling node set.
7. A customer mining device, comprising:
the graph neural network construction module is used for acquiring customer information and product information and constructing a graph neural network according to the customer information and the product information, and nodes in the graph neural network are customers;
the characteristic acquisition module is used for extracting node characteristics of all nodes in the graph neural network through the node characteristic extraction model;
the node pooling module is used for calculating the similarity of the nodes according to the node characteristics, combining the nodes with the similarity reaching a preset standard into one node and obtaining a pooled node set;
the monitoring processing module is used for installing probes in the pooling node set, the probes are used for monitoring the nodes in the pooling node set in real time, obtaining the node states of the nodes in the pooling node set, and presetting the nodes according to the node states;
the model building module is used for selecting partial nodes in the graph neural network to form a training set; extracting the characteristics of the nodes in the training set to obtain characteristic vectors; reconstructing the characteristic vector through a decoder to obtain the input characteristics of the nodes in the training set; and adjusting model parameters to enable the deviation between the input features and the feature vectors to be lower than a preset standard, so as to obtain a node feature extraction model.
8. The apparatus of claim 7, wherein the graph neural network construction module is specifically configured to:
acquiring customer information and product information, and constructing a customer-product graph and a customer-customer graph;
and combining the client and product graph and the client and client graph to construct a graph neural network.
9. The apparatus of claim 7, wherein the node pooling module is specifically configured to:
after the node characteristics are obtained, two nodes are selected each time, and the similarity of the two nodes is calculated by utilizing the cosine of the included angle.
10. The apparatus of claim 7, wherein the monitoring processing module is specifically configured to:
when the node state is logged off, the preset treatment is to delete the logged-off node from the graph neural network;
when the node state is inactive, the preset processing is to place the inactive node into a region to be observed of the graph neural network;
when a node and a new node have a transaction, the preset processing is to record the transaction times among the nodes, when the transaction times reach the preset value, the new node is added into the neighbor node of the node, and the new node is marked as a new node.
11. The apparatus of claim 7, wherein the model building module is specifically configured to:
selecting a node as a target node, finding out a neighbor node of the node, and recursively finding out the neighbor node of the neighbor node until the neighbor node recurses to the edge of a set;
drawing a high-order subgraph of the target node, and sequentially extracting the characteristics of the nodes in each layer of subgraph;
weighting and aggregating the characteristics of the nodes in the high-order subgraph layer by layer to obtain the node characteristics of a target node;
and splicing the node characteristics with the node characteristics of the neighbor nodes of the target node, and processing the spliced node characteristics by using the weight matrix and the nonlinear function to obtain the characteristic vector of the target node.
12. The apparatus of claim 7, further comprising a product recommendation module specifically configured to:
when the nodes are marked as new, analyzing the customer information marked as the new nodes to obtain customer characteristics;
and according to the characteristics of the customers, searching and recommending a product list matched with the newly added node in the pooling node set.
13. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the method of any one of claims 1 to 6 when executing the computer program.
14. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program which, when executed by a processor, implements the method of any one of claims 1 to 6.
15. A computer program product, characterized in that the computer program product comprises a computer program which, when being executed by a processor, carries out the method of any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211528914.6A CN115730147A (en) | 2022-11-30 | 2022-11-30 | Customer mining method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211528914.6A CN115730147A (en) | 2022-11-30 | 2022-11-30 | Customer mining method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115730147A true CN115730147A (en) | 2023-03-03 |
Family
ID=85299711
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211528914.6A Pending CN115730147A (en) | 2022-11-30 | 2022-11-30 | Customer mining method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115730147A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116453141A (en) * | 2023-06-13 | 2023-07-18 | 平安银行股份有限公司 | Identification method and device for bill latent passenger and electronic equipment |
-
2022
- 2022-11-30 CN CN202211528914.6A patent/CN115730147A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116453141A (en) * | 2023-06-13 | 2023-07-18 | 平安银行股份有限公司 | Identification method and device for bill latent passenger and electronic equipment |
| CN116453141B (en) * | 2023-06-13 | 2023-10-13 | 平安银行股份有限公司 | Identification method and device for bill latent passenger and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11074237B2 (en) | Method and system to estimate the cardinality of sets and set operation results from single and multiple HyperLogLog sketches | |
| Bao et al. | Autoconfig: Automatic configuration tuning for distributed message systems | |
| EP3493109A1 (en) | System and methods for faster processor comparisons of visual graph features | |
| CN104391879B (en) | The method and device of hierarchical clustering | |
| CN114742477B (en) | Enterprise order data processing method, device, equipment and storage medium | |
| CN115730147A (en) | Customer mining method and device | |
| JP2023546040A (en) | Data processing methods, devices, electronic devices, and computer programs | |
| CN111738331A (en) | User classification method and device, computer-readable storage medium and electronic device | |
| CN115730947A (en) | Bank customer loss prediction method and device | |
| CN113641906A (en) | System, method, device, processor and medium for realizing similar target person identification processing based on fund transaction relation data | |
| CN113204692A (en) | Method and device for monitoring execution progress of data processing task | |
| CN110727670A (en) | Data structure prediction transfer and automatic data processing method based on flow chart | |
| CN112633100B (en) | Behavior recognition method, behavior recognition device, electronic equipment and storage medium | |
| CN110377525B (en) | Parallel program performance prediction system based on runtime characteristics and machine learning | |
| CN111367971A (en) | Financial system abnormity auxiliary analysis method and device based on data mining | |
| CN113946717A (en) | Sub-map index feature obtaining method, device, equipment and storage medium | |
| CN113298120B (en) | Fusion model-based user risk prediction method, system and computer equipment | |
| CN114565033A (en) | Method and system for constructing characteristic parameter system of energy storage operation condition and storage medium | |
| CN114155410A (en) | Graph pooling, classification model training and reconstruction model training method and device | |
| CN103425476B (en) | For providing the method and system of data for the user interface of performance monitoring | |
| CN113610225A (en) | Quality evaluation model training method and device, electronic equipment and storage medium | |
| CN111090554A (en) | Method, device, equipment and storage medium for predicting server performance test time | |
| CN112750047A (en) | Behavior relation information extraction method and device, storage medium and electronic equipment | |
| Natalino et al. | Machine-learning-as-a-service for optical network automation | |
| Halici et al. | Object localization without bounding box information using generative adversarial reinforcement learning |
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 |