CN115186099A - Marketing and marketing method and system based on multi-dimensional construction knowledge graph - Google Patents
Marketing and marketing method and system based on multi-dimensional construction knowledge graph Download PDFInfo
- Publication number
- CN115186099A CN115186099A CN202210610838.7A CN202210610838A CN115186099A CN 115186099 A CN115186099 A CN 115186099A CN 202210610838 A CN202210610838 A CN 202210610838A CN 115186099 A CN115186099 A CN 115186099A
- Authority
- CN
- China
- Prior art keywords
- enterprise
- graph
- knowledge
- marketing
- relationship
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/284—Relational databases
- G06F16/288—Entity relationship models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/901—Indexing; Data structures therefor; Storage structures
- G06F16/9024—Graphs; Linked lists
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/02—Knowledge representation; Symbolic representation
-
- 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
-
- 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
-
- 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
- G06Q40/12—Accounting
- G06Q40/125—Finance or payroll
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- General Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- Development Economics (AREA)
- Strategic Management (AREA)
- Data Mining & Analysis (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Economics (AREA)
- Technology Law (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Animal Behavior & Ethology (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a marketing and marketing method and system based on a multi-dimensional constructed knowledge graph, belonging to the field of artificial intelligence, and the method comprises the following steps: analyzing enterprise relations, extracting entities, relations and attributes, constructing a plurality of enterprise knowledge maps and storing the enterprise knowledge maps to obtain a first map database; calculating to obtain graph calculation data characteristics in the first graph database; establishing an industry chain relation according to the enterprise knowledge graph and the graph calculation data characteristics to generate an industry knowledge graph; based on the enterprise knowledge map and the industrial knowledge map, matching financial statement data to identify financing requirements of clients on and off the enterprise; and displaying the enterprise knowledge map and the industrial knowledge map, and displaying and providing a query recognition function. The method can be used for identifying the upstream and downstream relationship of the industrial chain and supporting the development of customers of the industrial chain, thereby realizing the public marketing.
Description
Technical Field
The invention relates to the field of artificial intelligence, in particular to a marketing and marketing method and system based on a multi-dimensional constructed knowledge graph.
Background
With the continuous development and progress of the technology, the bank field also continuously applies new technology to the corresponding field, and the marketing field of public customers also needs to use new technical means, so that the marketing and role analysis personnel can use the marketing and role analysis technology conveniently, the customer relationship can be better constructed, the customer association relationship can be shown, and the deep insight of the customers can be supported, wherein the deep insight comprises group relationship, investment relationship, public and private relationship, guarantee relationship and the like. The new technology can also be applied to the aspect of portrait analysis, and the deep insights including the analysis of actual controllers, actual beneficiaries, key path exploration among enterprises, enterprise marketing events, enterprise risk events, enterprise portrait and the like of the enterprises can be performed from industry to enterprise, from single to related and from static to dynamic.
The knowledge graph can well solve the problems, the knowledge graph and a related algorithm are used for extracting entity relationship attributes of different relationships among enterprises and the like to form knowledge fusion to construct a corresponding knowledge graph model, a relationship network is generated by combining a corresponding graph calculation method to well reflect the relationships among the enterprises, a non-simple linear relationship of the generated relationships among the enterprises is constructed, influence is contained in the linear relationship, affinity calculation is carried out, and business personnel can explore business requirements of different relationships among the upstream and downstream of the enterprises and the like based on the relationship network displayed by graph analysis. Background data of the map can be stored through the map database, efficient query is carried out, the established relation map is displayed at the front end, analysis and use of marketers can be facilitated, and achievements of the knowledge map are applied to industrial chain relation establishment and upstream and downstream financing demand marketing. Therefore, it is desirable to provide a knowledge-graph-based marketing method and system.
Disclosure of Invention
In order to achieve the purpose, the technical scheme of the application provides a marketing and marketing method and system based on a multi-dimensional constructed knowledge graph. According to the requirements, a knowledge graph method is introduced into the public marketing field, and according to the research of the public business field, the relationship network of an enterprise is constructed into a plurality of (7) different enterprise relationship graphs, wherein the relationship graphs comprise a fund link relationship, an enterprise stock control relationship, a stockholder association relationship, a stockholder relative relationship, an enterprise debt relationship, a guarantee relationship and a relationship graph of a quota occupation relationship, and the relationship graphs are used as a basis for product requirements and enterprise default identification based on the relationship graphs, wherein the guarantee relationship and the quota occupation relationship are not commonly used for constructing the relationship graphs among public enterprises, and by quoting the two relationships, risks can be observed through the guarantee relationship, and business opportunities, businesses or identification risks can be observed through the quota occupation relationship.
According to a first aspect of the technical scheme of the application, a marketing-marketing method based on multi-dimensional construction of a knowledge graph is provided, and the method comprises the following steps:
s1: constructing an enterprise knowledge graph: analyzing enterprise relations, extracting entities, relations and attributes, constructing a plurality of enterprise knowledge maps and storing the enterprise knowledge maps to obtain a first map database;
s2: graph calculation: calculating to obtain graph calculation data characteristics in the first graph database;
s3: and (3) industrial knowledge map construction: establishing an industry chain relation according to the enterprise knowledge graph and the graph calculation data characteristics to generate an industry knowledge graph;
s4: identifying financing requirements: based on the enterprise knowledge map and the industrial knowledge map, matching financial statement data to identify financing requirements of clients on and off the enterprise;
s5: and (3) knowledge graph display: and displaying the enterprise knowledge graph and the industrial knowledge graph, and displaying and providing a query recognition function.
Further, in step S1, the enterprise relationship includes: the method comprises the steps of inter-enterprise fund link relationship, enterprise stock control relationship, shareholder associated enterprise relationship, shareholder relative associated enterprise relationship, enterprise creditor relationship, enterprise guarantee relationship and credit line occupation relationship.
Further, in the step S1, the shareholder associated business relationship includes: stockholder information, joint occupational association relationship, stockholder association business relationship of the enterprise, conditions of the enterprise associating joint occupational and director of the enterprise, prisoner, senior manager/natural person stockholder/legal person holding stocks externally, and information of senior manager, natural person stockholder, legal person of the enterprise.
Further, in the step S1, the shareholder relatives associating business relationships includes: enterprise shareholder relativity, high-level managers to the enterprise, natural person shareholders, corporate relativity associations, and enterprise relations of common duties and external holdings of high-level managers/natural person shareholders/corporate.
Further, in the step S1, the enterprise security relationship includes a security relationship between enterprises, and security relationships such as a simple security, a cyclic security, a financing security, and a platform security are identified.
Further, in the step S1, the credit line occupation relationship includes an inter-enterprise line occupation relationship and an inter-supply chain enterprise, group client, bank enterprise and other entities line occupation relationship.
Further, the step S1 specifically includes:
s1.1: analyzing the relationship of fund links among enterprises, the relationship of enterprise stock control, the relationship of shareholder related enterprises, the relationship of shareholder relative related enterprises, the relationship of enterprise debt and authority, the relationship of enterprise guarantee and the relationship of credit line occupation;
s1.2: and respectively performing entity extraction, relation extraction and attribute extraction on the information in the relation, performing knowledge depth fusion and knowledge combination on the extracted entities, and respectively constructing an enterprise knowledge map to obtain a first map database.
Further, in the step S1.1, in the process of analyzing the relationship of the fund chain between the enterprises, the fund relationship distinguishes transaction directions (i.e. upstream and downstream), and meanwhile, the fund relationship (without the stock right relationship) is added, and an enterprise marked with the stock right relationship in the fund relationship-related enterprises needs to determine with business personnel whether the fund relationship should be removed.
Further, in step S2, the graph calculation data characteristics include a node importance value, a correlation index, a similarity value, loop information, and shortest path information.
Further, the step S2 specifically includes:
s2.1: calculating importance values of all nodes under different graph relations by using a centrality algorithm;
s2.2: carrying out community division by using a community discovery algorithm to obtain correlation indexes among nodes under different relationships;
s2.3: calculating a similarity value between two different entity vertexes by using a similarity algorithm;
s2.4: outputting loop information and shortest path information;
s2.5: storing the graph-computation-data features to the first graph database.
Further, the centrality algorithm in step S2.1 includes algorithms such as medium centrality, approach centrality, degree centrality, harmonic centrality, page rank, medium centrality, and the like.
Further, the community discovery algorithm in step S2.2 includes algorithms such as the venturi algorithm, label propagation, strongly connected component, weakly connected component, and trigonometric statistic.
Further, the similarity calculation method in step S2.3 includes algorithms such as cosine similarity, euclidean distance, and Jaccard similarity.
Further, the step S3 specifically includes:
s3.1: analyzing the enterprise knowledge map and the graph calculation data characteristics to generate a supply chain relation and establish a supply chain knowledge base, and generating an industry knowledge map by combining the out-of-line relation data and the industry chain relation;
s3.2: judging the authenticity of the supply chain relation through a business rule or an expert rule, and eliminating an invalid supply chain relation;
s3.3: and (4) combining all the industry knowledge maps, displaying and providing a query recognition function.
Further, in the step S3.2, the method for determining authenticity of the supply chain relationship includes selecting an industry enterprise customer with significant characteristics of an industry chain, mining a transaction opponent (a fund chain relationship, a transaction mode including a bill and the like) of the industry enterprise customer as a potential upstream supplier and a potential downstream customer, and determining whether there is a business relationship by combining the relationship of the transaction and the like and the enterprise attribute characteristics, thereby identifying the supply chain relationship customer.
Further, the step S4 specifically includes:
s4.1: based on the enterprise knowledge graph and the industrial knowledge graph, combining the graph calculation data characteristics to extract characteristic information of enterprise operation data, and analyzing client characteristics of upstream and downstream clients of the enterprise;
s4.2: the method comprises the steps of identifying potential marketing opportunities or loss signals through the enterprise knowledge graph, the industrial knowledge graph and state changes of the enterprise knowledge graph and the industrial knowledge graph in a certain period, analyzing historical longitudinal comparison analysis of asset liability rate, liquidity and accounts receivable in enterprise financial report data of products with financing requirements according to client characteristics of upstream and downstream clients of an enterprise, simultaneously comparing and analyzing the historical longitudinal comparison analysis with the enterprises of the same industry and the same type, generating a marketing list by combining enterprise client subdivision, and determining financing requirements of upstream and downstream clients of the enterprise.
Further, in the step 4.1, the client characteristics of the upstream and downstream clients of the enterprise include financial characteristics of the enterprise: accounts receivable, bills receivable, cash flow, etc.; the relationship characteristic of the fund chain: transaction amount, transaction frequency, transaction mode, number of opponents and the like.
According to a second aspect of the invention, there is provided a marketing and marketing apparatus for building a knowledge graph based on multiple dimensions, the apparatus operating based on the method provided in any one of the preceding aspects, the apparatus comprising:
the enterprise knowledge graph construction unit is used for analyzing enterprise relations, extracting entities, relations and attributes, constructing a plurality of enterprise knowledge graphs and storing the enterprise knowledge graphs to obtain a first graph database;
a graph calculation unit for calculating a graph calculation data feature in the first graph database;
the industry knowledge graph construction unit is used for constructing an industry chain relation according to the enterprise knowledge graph and the graph calculation data characteristics to generate an industry knowledge graph;
the financing demand identification unit is used for identifying the financing demand of the client on the upstream and the downstream of the enterprise based on the enterprise knowledge map and the industrial knowledge map in cooperation with financial statement data;
and the knowledge map display unit is used for displaying the enterprise knowledge map and the industrial knowledge map, and displaying and providing a query and identification function.
According to a third aspect of the present invention, there is provided a marketing-to-public system for building a knowledge graph based on multiple dimensions, the system comprising: a processor and a memory for storing executable instructions; wherein the processor is configured to execute the executable instructions to perform the method for marketing to public based on multi-dimensional construction of a knowledge graph according to any one of the above aspects.
According to a fourth aspect of the present invention, there is provided a computer-readable storage medium, on which a computer program is stored, which, when executed by a processor, implements the method for marketing on the basis of multidimensional construction of a knowledge graph according to any one of the above aspects.
The invention has the following beneficial effects:
according to the technical scheme, the core enterprises in the supply chain network can be found out better through the centrality algorithm, so that the demands of the bank for judging services such as financing and the like are facilitated, and the judgment success rate of the core enterprises is improved. Community discovery algorithms can help identify companies with the same characteristics that can belong to the same category, providing a near financing service. The similarity algorithm helps to calculate the distances between different nodes and judge the similarity between the nodes, and the nodes with high similarity can execute the same rules and models to provide approximate services to meet the requirements. The obtained graph variables are put into a subsequent financing recommendation model, and the accuracy and the promotion degree of the model are effectively improved.
The invention introduces the knowledge graph and graph database in the marketing field, the graph database can facilitate analysts to inquire the multi-dimensional relationship of the same object, and the inquiry efficiency is far higher than that of the traditional relational database. The knowledge graph can be conveniently used by people marketing and analyzing roles, the behaviors of existing customers in a virtual network or a small group can be better known at the front end by extracting entities, attributes and relationships and combining graph calculation and storage modes, the overall view of the relationships among the customers can be conveniently established, and the knowledge graph can be better used for guiding applications such as social network marketing of enterprises.
The invention introduces a plurality of different constructions of knowledge maps for public relations, wherein the two categories of fund link relations, enterprise stock control relations, shareholder association relations, shareholder relative enterprise relations and enterprise creditor relations are the fields of common public business analysis, the newly-introduced guarantee relations and credit occupation relations are not commonly used for the client relations in the public marketing field, the risks can be observed through the guarantee relations by identifying different guarantee relations, the credit occupation relations among enterprises can be identified, the business opportunities can be observed, the business can be developed or identified, 5 relations such as the fund link can be further supplemented, the associated enterprises comprise two categories, one is common duties, the other is the enterprises which supervise the external holdings of the shareholder/legal owners of natural people and the relatives thereof, and the transaction directions (namely upstream and downstream) can be distinguished for the fund relation data
According to the invention, the established knowledge maps with various relationships are used for extracting the industrial chain relationship and upstream and downstream fund relationship among enterprises, so that accurate marketing is conveniently carried out on company clients, a marketing thought taking clients as centers is realized, and according to specific company business requirements and consultation design schemes, an identification model of the upstream and downstream relationship of the industrial chain is developed and completed, and the construction of the industrial chain is completed.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.
FIG. 1 is a diagram illustrating a completed knowledge-graph structure constructed in accordance with an embodiment of the present invention;
FIG. 2 is a logic diagram of the system module operation completed according to the embodiment of the present invention;
fig. 3 is a flow chart of the work flow completed by the construction according to the embodiment of the invention.
Detailed Description
The present invention is described in detail below with reference to examples, it should be noted that the examples are only for illustrative purposes and should not be construed as limiting the scope of the present invention, and that those skilled in the art may make insubstantial modifications and adaptations to the invention described above.
In order to make the aforementioned objects, features and advantages of the present invention more comprehensible, the present invention is described in detail with reference to the accompanying drawings and the detailed description thereof.
The technical scheme of the invention provides a method and a system for marketing based on a multidimensional constructed knowledge graph. The knowledge graph is stored by using the graph database, so that on one hand, the problem of efficiency of calculating the relation between multi-dimensional enterprises by the traditional database can be solved, on the other hand, the relation between the relation and the distant and distant of various relations can be calculated by a graph algorithm, the upstream and downstream financial fund relation of the enterprise is established based on the mapped relation data, the fund demand of an enterprise user is identified, and the product demand and the enterprise default identification based on the relation graph are used as the basis. And extracting the entities, attributes and relationships from the relational graph background, and displaying the entities, attributes and relationships at the front end in a graph calculation and storage mode. And establishing a complete picture of the relationship among the clients, and guiding the social network marketing of the enterprise and other applications. And displaying the constructed different relationships to corresponding marketing personnel and business personnel more conveniently at the front end.
As shown in fig. 1, an industry chain relation is extracted through 7 constructed knowledge maps, an industry chain is constructed, a supply chain relation is identified according to an inline data algorithm, a fund transaction rule is analyzed, financial industry enterprise nodes are removed, a supply chain relation is generated by combining industry upstream and downstream labels, a supply chain relation is artificially compounded, and a supply chain knowledge base is established. Meanwhile, an industry map is constructed by combining the off-line relation data and the industry data, cross-industry chain analysis can be performed by means of the industry map, a target customer group is focused and drilled down to a customer map and a customer figure, and the target customer group is analyzed. And displaying an industrial chain relation map, identifying the upstream and downstream relation of the industrial chain, and supporting industrial chain extension.
Examples
As shown in fig. 2-3, the method specifically comprises the following steps:
s1: analyzing the 7 enterprise relations, extracting the entities, the relations and the attributes, and constructing the corresponding knowledge graph
S1.1: establishing a fund chain relation map among enterprises, analyzing fund relation traffic among the enterprises, performing entity extraction, relation extraction and attribute extraction on information in the fund relation, further performing knowledge fusion on extracted entities to form entity links, performing deep fusion and combination on obtained knowledge, establishing a knowledge map body, storing the knowledge map body in a map database, distinguishing transaction directions (namely upstream and downstream) for the fund relation, simultaneously increasing the fund relation (not containing a share relationship), and removing enterprises with share association relation from the fund relation association enterprises;
s1.2: establishing an enterprise stock control relation map, analyzing the stock control relation among enterprises, performing entity extraction, relation extraction and attribute extraction on information in the stock control relation, further performing knowledge fusion on the extracted entities to form entity links, performing deep fusion and knowledge combination on the obtained knowledge, and establishing a knowledge map body to store in a map database;
s1.3: establishing a relationship map of shareholder-related enterprises, analyzing information of shareholder of the enterprises, co-duties association relationship, shareholder-related enterprise relationship, analyzing co-duties and directors of the related enterprises, supervising affairs, the method comprises the following steps that (1) information of a senior manager (hereinafter referred to as board prisoner)/natural person stockholders/legal persons is extracted under the condition that the legal persons hold stocks of an enterprise, and entity relations of the board prisoner, the natural person stockholders and the legal persons are constructed to form a knowledge graph ontology which is stored in a graph database;
s1.4: establishing a shareholder relative enterprise relationship map, analyzing the shareholder relative relationship of the enterprise, performing data extraction and data integration of a knowledge map on information of the enterprise relationship of the donor/the shareholder of the natural shareholder, the lawler relative relationship of the enterprise, the joint duties of the donor/the natural shareholder/the lawler, and establishing the knowledge map to be stored in a map database;
s1.5: establishing an enterprise debt-right relationship map, analyzing debt data among enterprises, performing information extraction and data integration on the debt information among the enterprises to obtain responsive entity relationship information, and further establishing a corresponding knowledge map of the debt-right relationship to be stored in a map database;
s1.6: establishing an enterprise guarantee relationship map, analyzing guarantee relationships among enterprises, identifying simple guarantees, circulating guarantees, financing guarantees, platform guarantees and other guarantee relationships, performing entity extraction, relationship extraction and attribute extraction on obtained information, further performing knowledge fusion on extracted entities to form entity links, performing deep fusion and knowledge combination on obtained knowledge, establishing a knowledge map body, and storing the knowledge map body in a map database;
s1.7: establishing a credit line occupation relation map, analyzing the line occupation relation among enterprises, and the line occupation relation among entities such as supply chain enterprises, group clients, bank enterprises and the like, establishing a knowledge map through line occupation information, and arranging to obtain corresponding data and storing the data into a map database.
S2: in the graph database, by using a graph algorithm to perform graph calculation on various relations such as fund relations, stockholder relations, debt-right relations and the like among enterprises, all relations can be combined together to perform corresponding calculation, and the importance of each node under different relations in the graph structure is obtained through calculation, wherein the importance of each node comprises degree centrality, intermediary centrality, proximity centrality and influence degree (PageRank page ranking) helps people to obtain important information by using a knowledge graph to help follow-up use on a public relation.
The medium centrality is as follows: the more times that one vertex acts as the shortest path between any two other vertices, the more times that one vertex acts as an intermediary, the larger the value of the mesocentrality is;
the approach to centrality: and calculating the shortest distance from all the vertexes of the whole graph to each vertex. The recenterness of a vertex is the inverse of the average of the shortest distances of the vertex to all vertices within the network;
centricity of degree: the number of neighbors of each vertex is calculated, and the neighbors can be used for finding out the most popular/friends in the social network;
harmonizing centrality: is a near-centrality variant, in order to better solve the problem of computing centrality in a disconnected graph;
page ranking: the algorithm calculates the importance of each vertex in the graph according to the number of the vertex pointed to the connected edges and the importance degree of the pointed to connected edges.
In addition, similarity indexes among nodes are calculated in the map by using a corresponding map algorithm, so that the constructed knowledge map is better utilized, wherein common algorithms such as:
the luwen algorithm also can become community partition algorithm, divide the node into different groups through the calculation to the calculation obtains the modularization index after grouping, and the higher the index is better that the community that means the division is better, through dividing the node into different communities, can all regard the node of different communities as belonging to the same type.
The label propagation can calculate propagation attenuation or enhancement trend of the indexes among different nodes, and by using the label propagation, the label indexes of different nodes can be assigned to adjacent nodes to judge the similarity of different nodes.
The strong connection component, the weak connection component and the triangular statistics can help people to divide different nodes into groups by using a graphical method and judge the correlation among the nodes.
Here, the similarity of the nodes may be obtained by calculating the distance between the nodes by directly using the similarity calculation method.
Obtaining a correlation index by the node importance and a graph algorithm, combining the correlation index with an up-and-down relationship, such as a fund chain relationship, namely the direction of fund, and an industry chain relationship, namely financial data of a supply chain, or analyzing by transaction data, such as that the fund transferring client is an upstream, the money receiving enterprise client is a downstream preliminary screening, and further judging whether the fund transferring client is the up-and-down relationship by combining exploration analysis of products held by the client and the like;
s2.1: and (4) calculating the intimacy indexes and influence degrees of different and combined relationships by using a centrality algorithm. Algorithms including medium centrality, near centrality, degree centrality, harmonic centrality, page rank, medium centrality, etc. may be used:
the center of the medium is as follows: the more times a vertex acts as the shortest path between any two other vertices, the more times a vertex acts as an "intermediary", where the larger the value of the intermediacy:
where dst represents the number of shortest paths from vertex s to vertex t, and dst () represents the number of nodes traversed in the shortest path from vertex s to vertex t.
The approach to centrality: calculating the shortest distance between all the vertexes of the whole graph and each vertex, wherein the approaching centricity of one vertex is the reciprocal of the average value of the shortest distances between the vertex and all the vertexes in the network
Where dis (i, j) represents the distance from node i to node j, g is the total number of nodes, and taking the reciprocal of the result indicates that the larger the value, the higher the proximity to the centrality.
Here, to consider the average length of the shortest path, it is necessary to normalize this score so that it represents the average length of the shortest path, not the sum of them. The normalized recentness calculation formula is as follows:
as can be seen from the above formula, if the shortest distances from a node to other nodes in the graph are all small, the node can be considered to have high recenterness. The average shortest distance to other nodes is the smallest, meaning that this node is located at the center of the graph from a geometric perspective.
Centrality in the degree: the number of neighbors per vertex is calculated, in the social network, to find the most popular/friends:
X ij whether the nodes i and j are directly connected or not is judged, the nodes i and j are connected to be 1 or not is judged, the degree centrality of the node i is calculated to be the total number of direct connection of the node i and other g-1 nodes, and as can be seen from the formula, the larger the network scale is, the higher the maximum possible value of the degree centrality is, and in order to eliminate the influence of the network scale change on the degree centrality, the standardization needs to be carried out:
in the normalized centrality measuring formula, the centrality value of the node i is divided by the maximum possible connection number of the other g-1 nodes to obtain the proportion of the network nodes directly connected with the node i, and the larger the proportion is, the higher the centrality is.
Page rank (PageRank): the algorithm calculates the importance of each vertex in the graph according to the number of the connected edges pointed to by a certain vertex and the importance degree pointed to by the connected edges:
where n is the number of all vertices, each vertex is given a PR value in advance (hereinafter PR value is used to refer to PageRank value), since PR value is physically the probability of being visited, where PR (T) is i ) i Is T i PageRank value of (C), L (T) i ) Is T i The number of outgoing links is the number of links (Degree) of the vertex, and the PageRank value of α is the accumulation of a series of page importance score values similar to T.
S2.2: the community discovery algorithm is used for calculating the relationship in the graph, and community division and graph structure calculation (clustering) are carried out.
The Luwen algorithm: the algorithm maximizes the modularity score of each community, the modularity is calculated based on the number of internal community links and the number of inter-community links, and for a network, the higher the modularity of the algorithm after the algorithm partitions the community, the better the algorithm is. The simplified calculation formula of the modularity of the divided single communities is as follows:
where L is the correlation coefficient in the whole group, L c Is the correlation coefficient, k, in a partition c Is the total number of nodes in a partition.
And (3) label propagation: the basic idea is to predict the labels of the nodes which are not marked by the labels of the nodes which are marked, the algorithm is simple and efficient, the algorithm is commonly used as a reference result of community division, and the defects that the result of each iteration has randomness and low accuracy are caused.
In a group of densely connected nodes, a single tag may soon dominate, but tags experience difficulty in traversing sparsely connected regions. Eventually, the labels get trapped in a set of densely connected vertices, and when the algorithm is complete, the vertices that eventually have the same label are considered part of the same community.
Strongly connected components: searching a set of the maximum connected components in the directed graph, wherein if directed edges which point to each other exist between any pair of vertexes in the set, the set is a strong connected component;
weakly connected components: connected vertex sets are found in the graph, where all vertices in the same set constitute a connected graph, typically used to analyze graph structures.
Performing triangular statistics: and counting the number of triangles in the neighborhood of each vertex in the graph, and calculating the clustering coefficient of the vertex to be used for quantifying the probability that connecting edges also exist between any neighbors for describing the vertex.
S2.3: the similarity between two different entity vertexes is calculated by using a similarity algorithm, and the greater the similarity is, the stronger some relevance exists between the two vertexes can be understood.
Cosine similarity: the neighborhood information of the vertex is used as the vector expression of the vertex, and the cosine similarity is calculated according to a cosine formula
Wherein A and B represent two vectors (A) 1 ，A 2 ，A 3 ...A n ),(B 1 ，B 2 ，B 3 ...B n )，A i ×B i The denominator is the length of vectors a and B, which are dot products of vectors a and B.
Euclidean distance: and (4) taking the neighbor information of the vertex as the vector representation of the vertex, and calculating the Euclidean similarity according to an Euclidean formula.
Jaccard similarity: set size of intersection of two vertex's neighbors/set size of union of two vertex's neighbors.
S2.4: the graph database galaxyBase supports the definition of subgraphs for searching a fixed mode, and is packaged into a storage process for being called by the query language cypher
And (3) finding a loop: defining that if a bidirectional path with indefinite length and reachable path exists between two vertexes, a loop exists between the two vertexes;
and (4) inquiring the shortest path: whether there is an associative path between two enterprises (shortest path between two vertices).
S2.5: and taking the structured data generated by the graph algorithm based on the graph data as new features, storing the new features into the existing feature library, enriching the composition of the feature library, and using the new features for training and applying the algorithm model of the subsequent upstream and downstream financing requirements.
S3: and constructing an industry chain relation according to the knowledge graph formed by the seven relations and a data result obtained by calculating the corresponding graph, generating an industry graph and identifying upstream and downstream enterprises.
S3.1: by analyzing the fund transaction rule of the existing seven map relations, enterprise nodes in the financial industry are removed, the upstream and downstream labels of the industry are combined to generate a supply chain relation, the supply chain relation is artificially compounded, and a supply chain knowledge base is established. Meanwhile, an industry map is constructed by combining the out-of-line relation data and the industry data;
s3.2: the cross-industry industrial chain analysis can be performed by means of the industrial map, the target customer group is focused and drilled down to the customer map and the customer portrait, and the target customer group is analyzed. The main functions of the industry map include: displaying an industrial chain relation map, identifying the upstream and downstream relation of an industrial chain, and supporting industrial chain extension;
s3.3: the method for identifying and inferring the supply chain relationship comprises the steps of selecting industry enterprise customers with obvious characteristics of an industry chain, excavating trade opponents (fund chain relationship, trade modes including bills and the like) of the industry enterprise customers as potential upstream suppliers and downstream customers, and judging whether business relationships exist or not by combining the relations of the trade and the like and enterprise attribute characteristics, so that the supply chain relationship customers are identified, and if the business relationships are false, the business relationships are directly eliminated; if the result is partial false, the correction can be carried out on site;
s3.4: the functions of industrial chain map, leading enterprise identification, industrial chain core enterprise identification, industrial chain member identification, industrial portrait, single enterprise map, detailed information exploration and the like are displayed through the industrial map generated by combining the database.
S4: based on the existing relational network data and the financial statement data, financing requirements of tourists and visitors of an enterprise are identified, a financial statement structure is optimized specifically for enterprise customers needing the requirements such as fund returns and the like, the customers' requirements are met by matching products such as blessing, warranty and the like, and customer value is improved.
S4.1: the characteristic information based on the relation graph and the enterprise operation data is refined and reprocessed, and the client characteristics of the upstream and downstream of the company client object are analyzed. The corresponding characteristics of processing comprise financial characteristics of enterprises, such as accounts receivable, bills receivable, cash flow and the like, and fund chain relation characteristics, such as transaction amount, transaction frequency, transaction mode, opponent number and the like;
s4.2: and identifying potential marketing opportunities or loss signals through the change of the incidence relation and the state, analyzing the longitudinal comparison and analysis of the assets liability rate, the mobility and the history of accounts receivable in the enterprise client financial report data of financing demand products, simultaneously comparing and analyzing the indexes with enterprises of the same industry and the same type, and generating a marketing list by combining client subdivision.
S5: the relation map is displayed, the map display and query functions are realized by depending on the existing functions of the galaxyBase map database, the registered address information of the client is combined with the corresponding positioning technology through the current relation network model of the client, and the registered address information is used as an output result in the relation network, so that the latest branch information can be displayed for business marketing personnel.
The relationship between enterprises is well embodied through the relationship network calculated through the graph algorithm, the relationship between the enterprises is a non-simple linear relationship, influence and intimacy calculation are included, and business personnel can explore business requirements between different relationships such as upstream and downstream of the enterprises based on the relationship network displayed by the graph analysis. And storing data through the background data graph, and displaying the relationship graph to related users through the front end of the graph database.
And an industrial chain map is constructed, and industrial chain map display, faucet enterprise identification, industrial chain core enterprise identification, industrial chain member identification, industrial portrait, single enterprise map and detailed information exploration are realized. The method comprises the steps of constructing a demand model such as company client financing based on an industrial chain, accurately marketing company clients through adaptive product types, realizing a client-centered marketing idea, and developing and completing the company client financing demand model according to specific company business demands and consultation design schemes.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrases "comprising a component of' 8230; \8230;" does not exclude the presence of another like element in a process, method, article, or apparatus that comprises the element.
The above-mentioned serial numbers of the embodiments of the present invention are only for description, and do not represent the advantages and disadvantages of the embodiments.
Through the above description of the embodiments, those skilled in the art will clearly understand that the above implementation method can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation method. Based on such understanding, the technical solutions of the present invention may be embodied in the form of a software product, which is stored in a storage medium (such as ROM/RAM, magnetic disk, optical disk) and includes instructions for enabling a terminal (such as a mobile phone, a computer, a server, an air conditioner, or a network device) to execute the method according to the embodiments of the present invention.
While the present invention has been described with reference to the particular illustrative embodiments, it is to be understood that the invention is not limited to the disclosed embodiments, but is intended to cover various modifications, equivalent arrangements, and equivalents thereof, which may be made by those skilled in the art without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (10)
1. A marketing and marketing method based on a multi-dimensional construction knowledge graph is characterized by comprising the following steps:
s1: establishing an enterprise knowledge graph: analyzing enterprise relations, extracting entities, relations and attributes, constructing a plurality of enterprise knowledge graphs and storing the enterprise knowledge graphs to obtain a first graph database;
s2: graph calculation: calculating to obtain graph calculation data characteristics in the first graph database;
s3: and (3) construction of an industrial knowledge map: establishing an industry chain relation according to the enterprise knowledge graph and the graph calculation data characteristics to generate an industry knowledge graph;
s4: financing demand identification: based on the enterprise knowledge map and the industrial knowledge map, matching financial statement data to identify financing requirements of clients on and off the enterprise;
s5: displaying a knowledge graph: and displaying the enterprise knowledge map and the industrial knowledge map, and displaying and providing a query recognition function.
2. The method for marketing to public based on multi-dimension construction of knowledge graph according to claim 1, wherein in the step S1, the enterprise relation comprises: the method comprises the steps of inter-enterprise fund link relationship, enterprise stock control relationship, shareholder associated enterprise relationship, shareholder relative associated enterprise relationship, enterprise creditor relationship, enterprise guarantee relationship and credit line occupation relationship.
3. The method for marketing public affairs based on multi-dimensional construction knowledge-graph according to claim 2, wherein the step S1 specifically comprises:
s1.1: analyzing the relationship of fund links among enterprises, the relationship of enterprise stock control, the relationship of shareholder related enterprises, the relationship of shareholder relative related enterprises, the relationship of enterprise debt and authority, the relationship of enterprise guarantee and the relationship of credit line occupation;
s1.2: and respectively performing entity extraction, relation extraction and attribute extraction on the information in the relation, performing knowledge depth fusion and knowledge combination on the extracted entities, and respectively constructing an enterprise knowledge map to obtain a first map database.
4. The method for marketing public based on multi-dimension construction of knowledge-graph according to claim 1, wherein in step S2, the graph calculation data characteristics include node importance value, correlation index, similarity value, loop information and shortest path information.
5. The marketing method for marketing based on multi-dimensional construction of knowledge graph according to claim 4, wherein the step S2 specifically comprises:
s2.1: calculating importance values of all nodes under different graph relations by using a centrality algorithm;
s2.2: carrying out community division by using a community discovery algorithm to obtain correlation indexes among nodes under different relations;
s2.3: calculating a similarity value between two different entity vertexes by using a similarity algorithm;
s2.4: outputting loop information and shortest path information;
s2.5: storing the graph computation data features to the first graph database.
6. The marketing method for marketing based on multi-dimensional construction of knowledge graph according to claim 1, wherein the step S3 specifically comprises:
s3.1: analyzing the enterprise knowledge map and the graph calculation data characteristics to generate a supply chain relation and establish a supply chain knowledge base, and generating an industry knowledge map by combining the out-of-line relation data and the industry chain relation;
s3.2: judging the authenticity of the supply chain relation through a business rule or an expert rule, and eliminating an invalid supply chain relation;
s3.3: and (4) combining all the industry knowledge maps, displaying and providing a query recognition function.
7. The marketing method for marketing based on multi-dimensional construction of knowledge graph according to claim 1, wherein the step S4 specifically comprises:
s4.1: extracting characteristic information of enterprise operation data by combining the graph calculation data characteristics based on the enterprise knowledge graph and the industrial knowledge graph, and analyzing the client characteristics of upstream and downstream clients of the enterprise;
s4.2: the method comprises the steps of identifying potential marketing opportunities or loss signals through the enterprise knowledge graph, the industrial knowledge graph and state changes of the enterprise knowledge graph and the industrial knowledge graph in a certain period, analyzing historical longitudinal comparison analysis of asset liability rate, liquidity and accounts receivable in enterprise financial report data of products with financing requirements according to client characteristics of upstream and downstream clients of an enterprise, simultaneously comparing and analyzing the historical longitudinal comparison analysis with the enterprises of the same industry and the same type, generating a marketing list by combining enterprise client subdivision, and determining financing requirements of upstream and downstream clients of the enterprise.
8. A marketing-to-business apparatus that constructs a knowledge graph based on multiple dimensions, the apparatus operating based on the method of any one of claims 1 to 7, the apparatus comprising:
the enterprise knowledge graph construction unit is used for analyzing enterprise relations, extracting entities, relations and attributes, constructing a plurality of enterprise knowledge graphs and storing the enterprise knowledge graphs to obtain a first graph database;
a graph calculation unit for calculating a graph calculation data feature in the first graph database;
the industry knowledge graph construction unit is used for constructing an industry chain relation according to the enterprise knowledge graph and the graph calculation data characteristics to generate an industry knowledge graph;
the financing demand identification unit is used for identifying financing demands of clients on the upstream and the downstream of the enterprise by matching financial statement data based on the enterprise knowledge map and the industrial knowledge map;
and the knowledge map display unit is used for displaying the enterprise knowledge map and the industrial knowledge map, and displaying and providing a query and identification function.
9. A marketing-to-public system based on multidimensional building of a knowledge graph, the system comprising: a processor and a memory for storing executable instructions; wherein the processor is configured to execute the executable instructions to perform the method of marketing to a multi-dimension based construction of a knowledgegraph according to any of claims 1 to 7.
10. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, implements the method for marketing on a multidimensional-based construction of a knowledge graph according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210610838.7A CN115186099A (en) | 2022-05-31 | 2022-05-31 | Marketing and marketing method and system based on multi-dimensional construction knowledge graph |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210610838.7A CN115186099A (en) | 2022-05-31 | 2022-05-31 | Marketing and marketing method and system based on multi-dimensional construction knowledge graph |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115186099A true CN115186099A (en) | 2022-10-14 |
Family
ID=83513399
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210610838.7A Pending CN115186099A (en) | 2022-05-31 | 2022-05-31 | Marketing and marketing method and system based on multi-dimensional construction knowledge graph |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115186099A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116452313A (en) * | 2023-06-14 | 2023-07-18 | 平安银行股份有限公司 | Method and device for calculating customer value in bank game customer group and electronic equipment |
| CN116579631A (en) * | 2023-07-11 | 2023-08-11 | 中国铁道科学研究院集团有限公司铁道建筑研究所 | Intelligent analysis method and system for preventing and controlling geological disasters in tunnel construction |
| CN116702899A (en) * | 2023-08-07 | 2023-09-05 | 上海银行股份有限公司 | Entity fusion method suitable for public and private linkage scene |
| CN116913460A (en) * | 2023-09-13 | 2023-10-20 | 福州市迈凯威信息技术有限公司 | Marketing business compliance judgment and analysis method for pharmaceutical instruments and inspection reagents |
-
2022
- 2022-05-31 CN CN202210610838.7A patent/CN115186099A/en active Pending
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116452313A (en) * | 2023-06-14 | 2023-07-18 | 平安银行股份有限公司 | Method and device for calculating customer value in bank game customer group and electronic equipment |
| CN116452313B (en) * | 2023-06-14 | 2023-09-19 | 平安银行股份有限公司 | Method and device for calculating customer value in bank game customer group and electronic equipment |
| CN116579631A (en) * | 2023-07-11 | 2023-08-11 | 中国铁道科学研究院集团有限公司铁道建筑研究所 | Intelligent analysis method and system for preventing and controlling geological disasters in tunnel construction |
| CN116702899A (en) * | 2023-08-07 | 2023-09-05 | 上海银行股份有限公司 | Entity fusion method suitable for public and private linkage scene |
| CN116702899B (en) * | 2023-08-07 | 2023-11-28 | 上海银行股份有限公司 | Entity fusion method suitable for public and private linkage scene |
| CN116913460A (en) * | 2023-09-13 | 2023-10-20 | 福州市迈凯威信息技术有限公司 | Marketing business compliance judgment and analysis method for pharmaceutical instruments and inspection reagents |
| CN116913460B (en) * | 2023-09-13 | 2023-12-29 | 福州市迈凯威信息技术有限公司 | Marketing business compliance judgment and analysis method for pharmaceutical instruments and inspection reagents |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110009174B (en) | Risk recognition model training method and device and server | |
| CN115186099A (en) | Marketing and marketing method and system based on multi-dimensional construction knowledge graph | |
| CN111476662A (en) | Anti-money laundering identification method and device | |
| CN109584048A (en) | The method and apparatus that risk rating is carried out to applicant based on artificial intelligence | |
| CN106844407B (en) | Tag network generation method and system based on data set correlation | |
| CN109785144A (en) | A kind of assets classes method, apparatus, equipment and medium | |
| CN113592568B (en) | Business opportunity mining method and device, computer equipment and storage medium | |
| CN104321794A (en) | A system and method using multi-dimensional rating to determine an entity's future commercial viability | |
| CN112699249B (en) | Knowledge graph-based information processing method, device, equipment and storage medium | |
| CN109840676B (en) | Big data-based wind control method and device, computer equipment and storage medium | |
| US11138269B1 (en) | Optimizing database query processes with supervised independent autonomy through a dynamically scaling matching and priority engine | |
| CN116402512B (en) | Account security check management method based on artificial intelligence | |
| CN110825817B (en) | Enterprise suspected association judgment method and system | |
| CN115456745A (en) | Small and micro enterprise portrait construction method and device | |
| CN116401379A (en) | Financial product data pushing method, device, equipment and storage medium | |
| Suresh et al. | Extracting of patterns using mining methods over damped window | |
| CN115545886A (en) | Overdue risk identification method, overdue risk identification device, overdue risk identification equipment and storage medium | |
| Verma et al. | Data mining: next generation challenges and futureDirections | |
| Zhao et al. | What is market talking about market-oriented prospect analysis for entrepreneur fundraising | |
| Rao et al. | BMSP-ML: big mart sales prediction using different machine learning techniques | |
| US11010675B1 (en) | Machine learning integration for a dynamically scaling matching and prioritization engine | |
| CN113989018A (en) | Risk management method, risk management device, electronic equipment and medium | |
| Kliestik | The application of mathematical modeling to predict the financial health of bussinesses | |
| Sathya et al. | Stock price prediction using reinforcement learning and feature extraction | |
| Olson et al. | Data mining process |
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 |