CN114372815A - Screening method for potential customers - Google Patents

Screening method for potential customers Download PDF

Info

Publication number
CN114372815A
CN114372815A CN202111401213.1A CN202111401213A CN114372815A CN 114372815 A CN114372815 A CN 114372815A CN 202111401213 A CN202111401213 A CN 202111401213A CN 114372815 A CN114372815 A CN 114372815A
Authority
CN
China
Prior art keywords
node
nodes
core
similarity
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
Application number
CN202111401213.1A
Other languages
Chinese (zh)
Inventor
唐欣蕾
季颖生
张政
沈佳辰
谢金慧
蔡明�
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.)
Chinaetek Service & Technology Co ltd
Original Assignee
Chinaetek Service & Technology Co ltd
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 Chinaetek Service & Technology Co ltd filed Critical Chinaetek Service & Technology Co ltd
Priority to CN202111401213.1A priority Critical patent/CN114372815A/en
Publication of CN114372815A publication Critical patent/CN114372815A/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
    • G06Q30/0203—Market surveys; Market polls
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
    • G06Q10/06311—Scheduling, planning or task assignment for a person or group
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
    • 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

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Human Resources & Organizations (AREA)
  • Strategic Management (AREA)
  • Development Economics (AREA)
  • Economics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Finance (AREA)
  • Accounting & Taxation (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • Marketing (AREA)
  • Theoretical Computer Science (AREA)
  • Game Theory and Decision Science (AREA)
  • Educational Administration (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • Data Mining & Analysis (AREA)
  • Technology Law (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The application provides a screening method of potential customers, which comprises the following steps: generating each node corresponding to each customer according to the account information of each customer; forming a connection path between two nodes corresponding to any two clients which are transacted in a preset time period and are in an upstream-downstream relationship in the same industrial chain to generate a directed graph; carrying out similarity matching on each first node connected with a core node corresponding to a core client in the directed graph and the core node to determine the similarity between each first node and the core node, and carrying out similarity matching on each second node connected with the same node in the directed graph and the core node to determine the similarity between each second node and the core node; and determining at least one target node from the first nodes and the second nodes according to the similarity so as to determine the client corresponding to the target node as a potential client. The screening method for the potential customers can improve the accuracy of the screened potential customers.

Description

Screening method for potential customers
Technical Field
The application relates to the technical field of computers, in particular to a screening method of potential customers.
Background
The digital upgrading process for public businesses, which are the most important assets and income sources of banks, is slow compared with the personal retail business of banks. At present, banks are highly homogeneous to public products, and competition patterns for public businesses are increasingly severe. With the promotion of general finance in China and the key support of small and medium-sized micro enterprises in policy, banks face massive public client groups, so that the key for improving the efficiency of public marketing is to effectively screen out potential clients of the banks from the massive public client groups.
In the related art, a specific scene and a target customer group can be described in a rule form by a rule-driven method, and a rating card or a customer figure and the like are formed, so that potential customers can be screened out according to the rating card or the customer figure and the like. Or machine learning can be adopted to mine mass data of the bank, so that the service scene and the target customer group are comprehensively analyzed, and potential customers are screened out.
However, since rule-driven approaches are limited by business experience, not all areas are well-documented. And the rules formed by weak rules or fuzzy conditions are often low in precision, and meanwhile, the rules may conflict with the rules, so that the screened potential customers are not accurate enough. Secondly, the rules are time-sensitive and require periodic adjustment, resulting in inefficient screening of public customers. And the mass data of the bank is mined by adopting machine learning, enough samples are needed to reflect the whole image of the object, and most of commercial banks have insufficient sample labels, so that the application difficulty of the supervised learning method is increased, and the accuracy of the screened potential customers is also reduced.
Disclosure of Invention
The embodiment of the application provides a screening method of potential customers, which improves the accuracy of screened potential customers while improving the determination efficiency of the potential customers.
In a first aspect, an embodiment of the present application provides a screening method for potential customers, including:
generating nodes corresponding to the customers one by one according to the account information of the customers;
forming a connection path between two nodes corresponding to any two clients which are transacted in a preset time period and are in an upstream-downstream relationship with each other in the same industrial chain to generate a directed graph;
carrying out similarity matching on each first node connected with a core node corresponding to a core client in the directed graph and the core node to determine the similarity between each first node and the core node, carrying out similarity matching on each second node connected with the same node in the directed graph and the core node to determine the similarity between each second node and the core node;
and determining at least one target node from each first node and each second node according to each similarity so as to determine the client corresponding to the target node as a potential client.
In one embodiment, performing similarity matching between each first node connected to a core node corresponding to a core client in the directed graph and the core node, and determining the similarity between each first node and the core node includes:
determining the similarity between each first node and a core node corresponding to the core customer according to the transaction scale ratio of each target customer and the core customer corresponding to each first node;
wherein the transaction size fraction is a ratio of a net worth of transactions of the core customer to the target customer to a sum of all net worth of transactions generated by the core customer.
In one embodiment, the net transaction value is determined according to a transaction difference value between the target customer and the core customer within the preset time period and a preset time weight.
In an embodiment, performing similarity matching between each second node in the directed graph, which is connected to the same node as the core node, and determining the similarity between each second node and the core node includes:
determining the similarity between each second node and the core node according to the same connection path occupation ratio between each second node and the core node;
the same connection path occupation ratio is a ratio of a total number of connection paths which are commonly flowed in and out by the second node and the core node in the directed graph to a total number of connection paths which are respectively flowed in and out by the second node and the core node in the directed graph.
In an embodiment, the upstream-downstream relationship between the customers is determined according to the upstream-downstream relationship between the industries corresponding to the customers.
In an embodiment, the determining at least one target node from each of the first nodes and each of the second nodes according to each of the similarities includes:
screening out nodes to be selected, of which the similarity of the second nodes is greater than a preset threshold value, from the second nodes;
and determining at least one target node from each first node and each node to be selected according to the similarity between each first node and each node to be selected and the core node.
In an embodiment, determining at least one target node from each of the first nodes and each of the candidate nodes according to a similarity between each of the first nodes and each of the candidate nodes and the core node includes:
connecting each node to be selected with the core node;
updating the attributes of the first nodes and the connection paths of the nodes to be selected and the core nodes according to the similarity of the first nodes and the similarity of the nodes to be selected and the core nodes, and acquiring a directed weighted graph;
calculating the directed weighted graph according to a PageRank webpage ranking algorithm to obtain the node weights of each first node and each second node to be selected in the directed weighted graph;
and determining at least one target node from each first node and each node to be selected according to the weight of each node, so as to determine the customer corresponding to the target node as a potential customer.
In an embodiment, the determining, according to the node weights, at least one target node from each first node and each candidate node to determine the customer corresponding to the target node as a potential customer includes:
sorting the first nodes and the nodes to be selected according to the node weights of the first nodes and the nodes to be selected to obtain the ranks of the first nodes and the nodes to be selected;
and obtaining the target nodes ranked before a preset rank from the first nodes and the nodes to be selected, so as to determine the clients corresponding to the target nodes as potential clients.
In one embodiment, the attribute of the connection path summarizes information for a transaction between two customers corresponding to the two connected nodes.
In one embodiment, the direction of the connection path is determined according to an upstream-downstream relationship between two clients corresponding to two of the nodes.
The screening method of potential customers provided by the embodiment of the application generates the directed graph according to the customers which are in the upstream and downstream relation with each other in the same industrial chain, and using the core node corresponding to the core client as the starting point, and using the connection path in the directed graph to obtain each first node in direct upstream and downstream relation with the core node, and the same node connected with the core node, namely, each second node sharing the same upstream and downstream nodes with the core node, and according to the similarity of each second node and each second node with the core node, the relationship between each second node and each second node with the core node is quantized, therefore, the method can utilize the connection path and the similarity to quickly search the client with strong correlation with the core client corresponding to the core node as the potential client, can screen out the potential client without providing a large number of samples and rules, and improves the determination efficiency of the potential client. And the potential customers with strong correlation with the core customers can be accurately determined by utilizing the connection paths and the similarity among the nodes of the directed graph, so that the accuracy of the screened potential customers is improved.
Drawings
In order to more clearly illustrate the technical solutions in the present application or the prior art, the drawings needed for the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
FIG. 1 is a schematic flow chart diagram illustrating a screening method for potential customers provided by an embodiment of the present application;
fig. 2 is a schematic diagram of a connection structure between a first node and a core node according to an embodiment of the present application;
fig. 3 is a schematic diagram of a connection structure between a second node and a core node according to an embodiment of the present application.
Detailed Description
To make the purpose, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
The embodiments of the present application will be described in detail below with reference to the accompanying drawings.
Fig. 1 is a schematic flow chart of a screening method for potential customers according to an embodiment of the present invention, and the method is applied to an electronic device, where the electronic device may specifically be a server or a terminal device, and is used to determine potential customers from a large number of public customers. As shown in fig. 1, the screening method for potential customers provided in this embodiment includes:
step 101, generating nodes corresponding to the clients one by one according to account information of the clients;
102, forming a connection path between two nodes corresponding to any two clients which are transacted in a preset time period and are in an upstream-downstream relationship with each other in the same industrial chain to generate a directed graph;
103, performing similarity matching on each first node connected with a core node corresponding to a core client in the directed graph and the core node to determine the similarity between each first node and the core node, and performing similarity matching on each second node connected with the same node in the directed graph and the core node to determine the similarity between each second node and the core node;
and 104, determining at least one target node from each first node and each second node according to each similarity, so as to determine the client corresponding to the target node as a potential client.
The method comprises the steps of generating a directed graph according to clients which are in upstream and downstream relation with each other in the same industrial chain, taking a core node corresponding to a core client as a starting point, obtaining each first node which is in direct upstream and downstream relation with the core node by utilizing a connection path in the directed graph, and obtaining each second node which is connected with the core node and has the same upstream and downstream nodes with the core node, and quantifying the relation between each second node and the core node according to the similarity between each second node and each core node, so that the clients which have strong correlation with the core client corresponding to the core node can be quickly searched by utilizing the connection path and the similarity as potential clients, the potential clients can be screened out without providing a large number of samples and rules, and the determination efficiency of the potential clients is improved. And the potential customers with strong correlation with the core customers can be accurately determined by utilizing the connection paths and the similarity among the nodes of the directed graph, so that the accuracy of the screened potential customers is improved.
In step 101, the account information of the customer may be bank-to-public account information, which includes customer identification, industry to which the customer belongs, related transaction, historical transaction records of the customer, transaction scale, and the like. And then generating a corresponding node by using the account information of the customer, and using the node to represent the account information of the customer.
In one embodiment, each customer is a customer in the same industry chain.
In step 102, after each node representing account information of each customer is generated, a connectivity graph G ═ V, E representing the upstream-downstream relationship between each customer is constructed from each node generated corresponding to each customer in the same industrial chain. Where V denotes a set of nodes, E denotes an edge set, and an edge denotes a connection path between two nodes. And the attribute of the connection path between any two connected nodes summarizes information for the transaction between two clients corresponding to the two connected nodes. The transaction summary information may specifically be a transaction net worth.
In one embodiment, the direction of the connection path is determined according to an upstream-downstream relationship between two clients corresponding to two of the nodes.
In an embodiment, the upstream-downstream relationship between the customers is determined according to the upstream-downstream relationship between the industries corresponding to the customers.
In one embodiment, for whether any two industries in the same industry chain are in an upstream-downstream relationship, the determination may be made by determining whether the industry correlation and the transaction direction between the two industries are centralized.
In one embodiment, the level of business relevance between any two industries may be determined based on the density of transactions between the two industries. Specifically, when the ratio of the total number of the customer pairs having transactions in the two industries to the number of the customers having transactions in the two industries in each customer is greater than a preset first threshold, it is determined that the correlation between the two industries is high. Wherein two clients with transactions constitute a client pair. The first threshold value can be set according to actual conditions.
In one embodiment, whether the transaction directions of any two industries are centralized or not can be determined according to the transaction direction concentration ratio of the two industries. Specifically, if there is a transaction between customers in two industries, more one of the two industries is selected as an industry transaction direction according to the net amount of the transaction greater than 0 or less than 0, and the ratio of the number of customer pairs in the direction to the total number of the customer pairs with the transaction is determined as the transaction direction concentration. And when the concentration ratio of the transaction directions is greater than a preset second threshold value, determining that the transaction directions between the two industries are concentrated. The second threshold value can be set according to actual conditions.
When the correlation between the two industries is high and the transaction directions are centralized, it is determined that the two industries have an upstream-downstream relationship, and the upstream-downstream direction is the transaction direction.
In one embodiment, each customer presets an industry label to which the customer belongs, and after the upstream and downstream relationship between the industries is determined, the industry to which each customer belongs can be identified according to the industry label to which each customer belongs, so that the upstream and downstream relationship between the customers can be determined according to the upstream and downstream relationship between the industries.
In one embodiment, a time window T is preset, such as a preset period of three months, half a year, or a year. If there is a transaction between two clients in the time window T and there is an upstream-downstream relationship in the industry chain, a directed edge from the top to the downstream is established between two nodes corresponding to the two clients. The edge attribute is the net value of the transaction, which may be negative depending on the direction.
Illustratively, a client a is upstream of a client B in the industry chain, and a transaction is generated between the client a and the client B within the time serial port T, so that a directed edge from the node 1 to the node 2 is established between the node 1 corresponding to the client a and the node 2 corresponding to the client B.
After all the nodes are connected according to the above conditions, a directed graph G ═ V, E can be obtained.
In step 103, after the directed graph is obtained, the radicalsFor a given core client list, the core node U corresponding to the core clientcAs a source point, searching and core nodes U in a directed graphcConnected, i.e. with the core node UcFirst nodes U in upstream and downstream relationship with each otherp。UcUpRepresenting a given core node UcCorresponding core client and first node UpDirect transaction transactions between corresponding target customers occur within the time window T as shown in fig. 2. After each first node is found, similarity matching is carried out on each first node and the core node, and the similarity between each first node and the core node is determined
Figure BDA0003365234650000081
In order to make the obtained similarity more accurate, in an embodiment, the relationship between the first node and the core node may be quantized to determine the similarity between the first node and the core node. Specifically, according to the transaction scale ratio of each target customer and core customer corresponding to each first node, determining the similarity between each first node and the core node corresponding to the core customer;
wherein the transaction size fraction is a ratio of a net worth of transactions of the core customer to the target customer to a sum of all net worth of transactions generated by the core customer.
In one embodiment, the net transaction value is determined according to a transaction difference value between the target customer and the core customer within the preset time period and a preset time weight.
For example, assuming a core client i, a target client j, and a time window T, the net value of a transaction between the core client and the target client may be estimated based on the transaction pipeline
Figure BDA0003365234650000082
Comprises the following steps:
Figure BDA0003365234650000091
the transaction difference value is greater than 0 to indicate that the fund flows out, and the transaction difference value is less than 0 to indicate that the fund flows in; ω (t) is a preset weight with respect to time, typically using a time decay function.
After the net value of the transaction between the core customer and the target customer is determined, the strength degree of the relationship between the core customer and the target customer can be judged according to the transaction scale ratio of the target customer and the core customer, and the strength degree of the relationship is used as the similarity between the first node corresponding to the target customer and the core node corresponding to the core customer.
The transaction scale ratio of the target customer to the core customer is as follows:
Figure BDA0003365234650000092
wherein, the numerator represents the net value of the transaction between the core client i and the target client j, and the denominator represents the sum of all the net values of the transaction occurring at the core client i; k denotes the customer with whom the transaction occurs with the core customer. The smaller the transaction size of the target customer occupying the core customer is, the weaker the business relationship between the upstream or downstream target customer and the core customer is, and the stronger the business relationship is. Thus, the trade size can be scaled
Figure BDA0003365234650000093
Similarity of core node corresponding to core client i as first node corresponding to target client j
Figure BDA0003365234650000094
Thereby accurately and effectively determining the similarity of the first node and the core node.
Meanwhile, the core node U corresponding to the core clientcAs a source point, searching and core nodes U in a directed graphcConnect the same node UmI.e. with the core node UcHaving a common upstream or downstream node UmEach second node Up′。UcUmUp' denotes a given core node UcCorresponding core guestThe user and the second node Up' enterprises with the same upstream or downstream suppliers between corresponding customers. As shown in fig. 3. After each second node is found, similarity matching is carried out on each second node and the core node, and the similarity between each second node and the core node is determined
Figure BDA0003365234650000095
Because the direct transaction does not exist between the customer corresponding to the second node and the core customer corresponding to the core node, the relationship between the second node and the core node is quantized, and the similarity between each second node and the core node can be determined according to the same connection path occupation ratio between each second node and the core node; the same connection path occupation ratio is a ratio of a total number of connection paths which are commonly flowed in and out by the second node and the core node in the directed graph to a total number of connection paths which are respectively flowed in and out by the second node and the core node in the directed graph.
For example, after each second node is obtained, the strength of the relationship between the second node and the core node may be determined according to the common associated node of the second node and the core node, that is, the ratio of the same connection paths between the second node and the core node is:
Figure BDA0003365234650000101
wherein the numerator represents the total number of paths commonly flowed in or out by the core node i and the second node j, and is determined by the total number of upstream nodes shared by the core node i and the second node j, and the total number of downstream nodes shared by the core node i and the second node j. The denominator represents the total number of all ingress and egress paths for core node i and second node j. e'ijThe binary relation indicating whether there is a directed edge from the core node i to the second node j is defined as follows:
Figure BDA0003365234650000102
because the ratio of the total number of the paths which are commonly flowed in or out of the core node i and the second node j to the total number of all the paths which are flowed in or out of the core node i and the second node j is smaller, the intersection between the two industries corresponding to the two nodes is smaller, namely the industry relationship between the two industries corresponding to the two nodes is weaker, and otherwise, the industry relationship is relatively stronger. Therefore, the same connection path can be used
Figure BDA0003365234650000103
As a second node j, similarity with a core node i
Figure BDA0003365234650000104
Thereby accurately and effectively determining the similarity of the second node and the core node.
In an embodiment, after determining the similarity between each first node and each second node and the core node, the first node or the second node corresponding to the similarity higher than a preset value may be used as a target node, and a customer corresponding to the target node is determined as a potential customer. The preset value can be set according to actual conditions.
Considering that the preset value is usually difficult to set, the first nodes and the second nodes can be sorted according to the similarity, and the nodes sorted at the top N bits are obtained as target nodes, wherein N is a positive integer.
However, since each second node is not directly connected to the core node, i.e. not directly related to the core node, each first node is directly related to the core node. Thus, in practice, even if the second node is more similar to the core node than the first node, the first node may still be more relevant to the core node than the second node. At this time, if the target nodes are determined by directly performing the sequencing according to the similarity between each first node and each second node and the core node, the situation that users corresponding to part of the target nodes are not potential users in the actual sense may occur, resulting in an inaccurate screening result. To this end, in an embodiment, the determining at least one target node from each of the first nodes and each of the second nodes according to each of the similarities includes:
screening out nodes to be selected, of which the similarity of the second nodes is greater than a preset threshold value, from the second nodes;
and determining at least one target node from each first node and each node to be selected according to the similarity between each first node and each node to be selected and the core node.
In an embodiment, after the similarity between each second node and the core node is obtained, according to a preset threshold, each node to be selected, of which the similarity is greater than the preset threshold, is first selected from the second nodes, and then each node to be selected and each first node are selected according to the similarity ranking, so that N target nodes are selected.
The correlation degree of each screened to-be-selected node and the core node is high enough by primarily screening each second node according to the preset threshold value, so that the correlation degree of the to-be-selected node and the core node is not weaker than that of the first node and the core node, at the moment, the target nodes are screened out from each to-be-selected node and each first node according to the similarity sorting, the correlation degree of the screened target nodes and the core node is also high enough, further, the potential users determined according to the target nodes are enabled to better meet the actual requirements, and the accuracy of the screened potential users is improved.
In order to further improve the accuracy of the screened potential users, in an embodiment, determining at least one target node from each of the first nodes and each of the candidate nodes according to the similarity between each of the first nodes and each of the candidate nodes and the core node includes:
connecting each node to be selected with the core node;
updating the attributes of the first nodes and the connection paths of the nodes to be selected and the core nodes according to the similarity of the first nodes and the similarity of the nodes to be selected and the core nodes, and acquiring a directed weighted graph;
calculating the directed weighted graph according to a PageRank webpage ranking algorithm to obtain the node weights of each first node and each second node to be selected in the directed weighted graph;
and determining at least one target node from each first node and each node to be selected according to the weight of each node, so as to determine the customer corresponding to the target node as a potential customer.
In an embodiment, after the nodes to be selected are screened out, a directed connection path from the core node to the candidate node is generated between each candidate node and the core node to update the directed graph, a potential client sub-graph formed by the core node, each first node and each candidate node with a connection path existing with the core node is extracted from the updated directed graph, and then the attribute of each connection path in the potential client sub-graph is updated to the similarity between the core node at the two ends of the path and the first node or the candidate node to obtain a directed weighted graph G ' ═ V ', E '. Wherein V represents a representation node subset consisting of the core node, each first node, and each candidate node, E' represents a directed edge, i.e., a set of connection paths, and the edge attribute represents the similarity obtained by relationship quantization.
After the directed weighted graph is obtained, the directed weighted graph is operated through a PageRank webpage ranking algorithm, which specifically comprises the following steps:
taking each first node and each candidate node as a one-degree neighbor node, and taking the initial weight a of the core node in the directed weighted graph0 (1)Set to 1, each one-degree neighbor node initial weight a0 (2)Set to 0, perform initialization.
After the initialization is completed, iterative recursive optimization is performed. And taking the core point as a current starting point, starting from the current starting point, transmitting the weight of the core point to each corresponding first-degree neighbor node according to the similarity corresponding to each connecting path, updating the node weights of the core node and each first-degree neighbor node according to different modes, combining the first-degree neighbor node and the current starting point set to be used as the starting point of the next iteration, and continuously updating the node weights of each first-degree neighbor node around the starting point.
For example, the ith step can accept objects according to different node weights in two types. Wherein, the node weight of the core node v can be updated as:
Figure BDA0003365234650000131
the first degree neighbor node v' may have its node weight updated as:
Figure BDA0003365234650000132
wherein D is a damping coefficient,
Figure BDA0003365234650000133
representing the similarity between two nodes.
And when the difference value between the updated node weight and the node weight before updating is smaller than a certain set threshold value, stopping iteration. Wherein, the set threshold value can be set according to the actual situation.
And after the iteration is stopped, outputting the node weight of each one-degree neighbor node, and selecting at least one-degree neighbor node as a target node according to the node weight of each one-degree neighbor node. For example, a one-degree neighbor node with a node weight greater than a preset weight is used as a target node, so that a customer corresponding to the target node is determined as a potential customer.
The setting of the preset weight is difficult considering that the setting needs to be dependent on experience. Therefore, in an embodiment, the first nodes and the nodes to be selected may be sorted according to the node weights of the first nodes and the nodes to be selected, so as to obtain the ranks of the first nodes and the nodes to be selected;
and obtaining the target nodes ranked before a preset rank from the first nodes and the nodes to be selected, so as to determine the clients corresponding to the target nodes as potential clients.
In one embodiment, all first-degree neighbor nodes composed of all first nodes and all candidate nodes are ranked from high to low according to the node weights of all first-degree neighbor nodes, then all first-degree neighbor nodes ranked in the top N positions are extracted and serve as target nodes, users corresponding to the target nodes are determined as potential users, therefore, the setting of preset weights is not needed, the situation that due to the fact that the preset weights are not accurately set, the target nodes are extracted too much or too little, even the target nodes cannot be extracted, the screening of the potential users is not accurate is avoided, and the screening efficiency for the potential users is improved.
Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions in the embodiments of the present application.

Claims (10)

1. A method for screening potential customers, comprising:
generating nodes corresponding to the customers one by one according to the account information of the customers;
forming a connection path between two nodes corresponding to any two clients which are transacted in a preset time period and are in an upstream-downstream relationship with each other in the same industrial chain to generate a directed graph;
carrying out similarity matching on each first node connected with a core node corresponding to a core client in the directed graph and the core node to determine the similarity between each first node and the core node, carrying out similarity matching on each second node connected with the same node in the directed graph and the core node to determine the similarity between each second node and the core node;
and determining at least one target node from each first node and each second node according to each similarity so as to determine the client corresponding to the target node as a potential client.
2. The method for screening potential customers according to claim 1, wherein performing similarity matching between each first node connected to a core node corresponding to a core customer in the directed graph and the core node to determine the similarity between each first node and the core node comprises:
determining the similarity between each first node and a core node corresponding to the core customer according to the transaction scale ratio of each target customer and the core customer corresponding to each first node;
wherein the transaction size fraction is a ratio of a net worth of transactions of the core customer to the target customer to a sum of all net worth of transactions generated by the core customer.
3. The method of claim 2, wherein the net transaction value is determined according to a difference between transactions of the target customer and the core customer within the predetermined time period and a predetermined time weight.
4. The method for screening potential customers according to claim 1 or 2, wherein performing similarity matching between each second node in the directed graph, which is connected to the same node as the core node, and determining the similarity between each second node and the core node comprises:
determining the similarity between each second node and the core node according to the same connection path occupation ratio between each second node and the core node;
the same connection path occupation ratio is a ratio of a total number of connection paths which are commonly flowed in and out by the second node and the core node in the directed graph to a total number of connection paths which are respectively flowed in and out by the second node and the core node in the directed graph.
5. The method for screening potential customers according to claim 1, wherein the upstream-downstream relationship between the customers is determined according to the upstream-downstream relationship between the industries corresponding to the customers.
6. The method for screening potential customers of claim 1, wherein the determining at least one target node from the first nodes and the second nodes according to the similarities comprises:
screening out nodes to be selected, of which the similarity of the second nodes is greater than a preset threshold value, from the second nodes;
and determining at least one target node from each first node and each node to be selected according to the similarity between each first node and each node to be selected and the core node.
7. The method for screening potential customers according to claim 6, wherein determining at least one target node from the first nodes and the candidate nodes according to the similarity between the core nodes and the candidate nodes, comprises:
connecting each node to be selected with the core node;
updating the attributes of the first nodes and the connection paths of the nodes to be selected and the core nodes according to the similarity of the first nodes and the similarity of the nodes to be selected and the core nodes, and acquiring a directed weighted graph;
calculating the directed weighted graph according to a PageRank webpage ranking algorithm to obtain the node weights of each first node and each second node to be selected in the directed weighted graph;
and determining at least one target node from each first node and each node to be selected according to the weight of each node, so as to determine the customer corresponding to the target node as a potential customer.
8. The method for screening potential customers according to claim 7, wherein the determining at least one target node from the first nodes and the candidate nodes according to the node weights to determine the customer corresponding to the target node as a potential customer comprises:
sorting the first nodes and the nodes to be selected according to the node weights of the first nodes and the nodes to be selected to obtain the ranks of the first nodes and the nodes to be selected;
and obtaining the target nodes ranked before a preset rank from the first nodes and the nodes to be selected, so as to determine the clients corresponding to the target nodes as potential clients.
9. The method of claim 1, wherein the attributes of the connection path are aggregated information for transactions between two customers corresponding to two connected nodes.
10. The method for screening potential customers of claim 9, wherein the direction of the connection path is determined according to an upstream-downstream relationship between two customers corresponding to two of the nodes.
CN202111401213.1A 2021-11-19 2021-11-19 Screening method for potential customers Pending CN114372815A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111401213.1A CN114372815A (en) 2021-11-19 2021-11-19 Screening method for potential customers

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111401213.1A CN114372815A (en) 2021-11-19 2021-11-19 Screening method for potential customers

Publications (1)

Publication Number Publication Date
CN114372815A true CN114372815A (en) 2022-04-19

Family

ID=81137884

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111401213.1A Pending CN114372815A (en) 2021-11-19 2021-11-19 Screening method for potential customers

Country Status (1)

Country Link
CN (1) CN114372815A (en)

Similar Documents

Publication Publication Date Title
CN110532542B (en) Invoice false invoice identification method and system based on positive case and unmarked learning
CN110390408B (en) Transaction object prediction method and device
CN112395487B (en) Information recommendation method and device, computer readable storage medium and electronic equipment
CN112561082A (en) Method, device, equipment and storage medium for generating model
CN111191825A (en) User default prediction method and device and electronic equipment
CN115545886A (en) Overdue risk identification method, overdue risk identification device, overdue risk identification equipment and storage medium
CN114691525A (en) Test case selection method and device
CN108304568B (en) Real estate public expectation big data processing method and system
CN111930944B (en) File label classification method and device
CN115049446A (en) Merchant identification method and device, electronic equipment and computer readable medium
CN112231299A (en) Method and device for dynamically adjusting feature library
CN115329207B (en) Intelligent sales information recommendation method and system
CN116502898A (en) Enterprise risk portrait generation method and device based on neural network
CN111582313A (en) Sample data generation method and device and electronic equipment
CN115730152A (en) Big data processing method and big data processing system based on user portrait analysis
CN113591881B (en) Intention recognition method and device based on model fusion, electronic equipment and medium
CN116204647A (en) Method and device for establishing target comparison learning model and text clustering
CN114372815A (en) Screening method for potential customers
CN114897607A (en) Data processing method and device for product resources, electronic equipment and storage medium
CN115034762A (en) Post recommendation method and device, storage medium, electronic equipment and product
Roelands et al. Classifying businesses by economic activity using web-based text mining
CN112632284A (en) Information extraction method and system for unlabeled text data set
CN110895564A (en) Potential customer data processing method and device
CN115187387B (en) Identification method and equipment for risk merchant
CN113656393B (en) Data processing method, device, electronic equipment and storage medium

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