A method and device for obtaining corporate behavior preferences

Abstract

The present invention provides a method and device for obtaining corporate behavior preferences, which can be used in the financial field or other technical fields. The method includes: constructing an enterprise knowledge graph based on the basic information and behavior information of each enterprise; calculating the centrality graph algorithm score of each enterprise under each relationship type based on the relationship type between the enterprises in the enterprise knowledge graph; inputting the basic information, behavior information and each centrality graph algorithm score of each enterprise into a pre-trained behavior preference prediction model to obtain the behavior preference of the enterprise, wherein the behavior preference prediction model is obtained after training based on the basic information, behavior information, each centrality graph algorithm score and the behavior preference labels predetermined by the sample enterprise. The device and electronic device are used to execute the above method. The method and device for obtaining corporate behavior preferences provided in the embodiments of the present invention can greatly improve the accuracy of determining corporate behavior preferences.

Description

Technical Field

[0001] The present invention relates to the financial field or other technical fields, and in particular to a method and device for acquiring corporate behavior preferences. Background Art

[0002] The expansion and maintenance of corporate customers (B-side) plays a vital role in the banking industry in terms of product marketing, C-side customer development, and capital accumulation. Finding the behavioral preferences of corporate customers and determining practical marketing paths are of great significance for customer retention and account expansion.

[0003] At present, in the process of building a prediction model for the behavior preference of corporate customer groups, it is only based on the mining of corporate information, such as corporate size, industry type, capital transaction volume, products held, credit financing information, etc., and is determined by adding statistical rules to the relational database. In fact, the behavior preference of an enterprise is not only obtained by simple statistics based on its own information. Only by dividing the behavioral characteristics of the corporate customer groups based on the basic situation of the enterprise itself, forming a one-sided marketing list, and only containing statistical introductions based on the basic information characteristics of customers, it is inevitable that the marketing path is unclear and the success rate is low. Summary of the invention

[0004] In view of the problems in the prior art, an embodiment of the present invention provides a method and device for acquiring enterprise behavior preferences, which can at least partially solve the problems in the prior art.

[0005] On the one hand, the present invention proposes a method for obtaining enterprise behavior preferences, comprising: constructing an enterprise knowledge graph based on basic information and behavior information of each enterprise, wherein the enterprise knowledge graph has each enterprise as a node and the relationship between each enterprise as an edge; calculating the centrality graph algorithm score of each enterprise under each relationship type based on the relationship type between each enterprise in the enterprise knowledge graph, wherein there is at least one relationship between two linked enterprises in the enterprise knowledge graph; inputting the basic information, behavior information and centrality graph algorithm scores of each enterprise into a pre-trained behavior preference prediction model to obtain the behavior preference of the enterprise, wherein the behavior preference prediction model is obtained after training based on the basic information, behavior information, centrality graph algorithm scores of sample enterprises and behavior preference labels predetermined by the sample enterprises.

[0006] Optionally, the calculating of the centrality graph algorithm score of each enterprise under each relationship type according to the relationship type between the enterprises in the enterprise knowledge graph includes: for each enterprise under each relationship type in the enterprise knowledge graph, calculating the centrality graph algorithm score of each enterprise using the following centrality graph algorithm score calculation formula: Among them, T1~T nare neighbor nodes that have a link relationship with node u, PR(u), PR(T1)...PR(T n ) represent nodes u, T1...T n The centrality graph algorithm score, C(T n ) is the node T n The degree refers to the number of neighboring nodes that the node is linked to; d is the damping coefficient, and (1-d) represents the probability that node u is linked to.

[0007] Optionally, after inputting the basic information, behavioral information and centrality graph algorithm scores of each enterprise into a pre-trained behavioral preference prediction model to obtain the behavioral preferences of the enterprise, the method also includes: determining a marketing path for the enterprise based on the knowledge graph shortest path algorithm according to the behavioral preference type of the enterprise.

[0008] Optionally, determining the marketing path for the enterprise based on the behavior preference type of the enterprise and based on the knowledge graph shortest path algorithm includes: obtaining the enterprise relationship type related to the behavior preference type of the enterprise in the enterprise knowledge graph; calculating the shortest path between the enterprise and each other enterprise under the relationship type in the enterprise knowledge graph based on the knowledge graph shortest path algorithm; and determining the marketing path for the enterprise based on the shortest path between each other enterprise and the enterprise.

[0009] Optionally, determining the marketing path for the enterprise based on the shortest paths between each other enterprise and the enterprise includes: obtaining the path depth of each of the shortest paths, and the centrality graph score of the key nodes in each of the shortest paths; and taking the shortest path in which the path depth is less than the target depth and the centrality graph score of the key nodes in the path is greater than the target score as the marketing path for the enterprise.

[0010] Optionally, after determining a marketing path for the enterprise based on a knowledge graph shortest path algorithm according to the behavioral preference type of the enterprise, the method further includes: displaying the marketing path in the form of a visualized enterprise knowledge graph.

[0011] Optionally, the behavior preference prediction model is trained based on the basic information, behavior information, centrality graph algorithm scores and predetermined behavior preference labels of the sample enterprises, including: constructing a sample enterprise knowledge graph according to the basic information and behavior information of each sample enterprise, wherein the sample enterprise knowledge graph has each sample enterprise as a node and the relationship between each sample enterprise as an edge; calculating the centrality graph algorithm score of each sample enterprise under each relationship type according to the relationship type between each sample enterprise in the sample enterprise knowledge graph, wherein there is at least one relationship between two sample enterprises linked in the sample enterprise knowledge graph; and training a behavior preference prediction model according to the basic information, behavior information, centrality graph algorithm scores and predetermined behavior preference labels of each sample enterprise.

[0012] On the other hand, the present invention provides a device for acquiring enterprise behavior preferences, comprising: a graph construction module, used to construct an enterprise knowledge graph based on the basic information and behavior information of each enterprise, wherein the enterprise knowledge graph has each enterprise as a node and the relationship between each enterprise as an edge; a calculation module, used to calculate the centrality graph algorithm score of each enterprise under each relationship type according to the relationship type between the enterprises in the enterprise knowledge graph, wherein there is at least one relationship between two linked enterprises in the enterprise knowledge graph; an acquisition module, used to input the basic information, behavior information and centrality graph algorithm scores of each enterprise into a pre-trained behavior preference prediction model to obtain the behavior preference of the enterprise, wherein the behavior preference prediction model is obtained after training based on the basic information, behavior information, centrality graph algorithm scores of sample enterprises and behavioral preference labels predetermined by the sample enterprises.

[0013] Optionally, the calculation module is specifically used to: for each enterprise under each relationship type in the enterprise knowledge graph, calculate the centrality graph algorithm score of each enterprise using the following centrality graph algorithm score calculation formula: Among them, T1~T n are neighbor nodes that have a link relationship with node u, PR(u), PR(T1)...PR(T n ) represent nodes u, T1...T n The centrality graph algorithm score, C(T n ) is the node T n The degree refers to the number of neighboring nodes that the node is linked to; d is the damping coefficient, and (1-d) represents the probability that node u is linked to.

[0014] Optionally, the device also includes: a determination module, used to determine a marketing path for the enterprise based on the knowledge graph shortest path algorithm according to the behavioral preference type of the enterprise.

[0015] Optionally, the determination module specifically includes: an acquisition unit, used to acquire the enterprise relationship type related to the behavioral preference type of the enterprise in the enterprise knowledge graph; a calculation unit, used to calculate the shortest path between each other enterprise under the relationship type in the enterprise knowledge graph and the enterprise based on the knowledge graph shortest path algorithm; a determination unit, used to determine the marketing path for the enterprise based on the shortest path between each other enterprise and the enterprise.

[0016] Optionally, the determination unit is specifically used to: obtain the path depth of each of the shortest paths, and the centrality graph score of the key nodes in each of the shortest paths; and use the shortest path in which the path depth is less than the target depth and the centrality graph score of the key nodes in the path is greater than the target score as the marketing path for the enterprise.

[0017] Optionally, the device also includes: a display module, used to display the marketing path in the form of a visual enterprise knowledge graph.

[0018] Optionally, the device also includes a model building module, which is used to: construct a sample enterprise knowledge graph based on basic information and behavioral information of each sample enterprise, wherein the sample enterprise knowledge graph has each sample enterprise as a node and the relationship between each sample enterprise as an edge; calculate the centrality graph algorithm score of each sample enterprise under each relationship type based on the relationship type between each sample enterprise in the sample enterprise knowledge graph, wherein there is at least one relationship between two sample enterprises linked in the sample enterprise knowledge graph; and train a behavior preference prediction model based on the basic information, behavioral information, centrality graph algorithm scores and predetermined behavior preference labels of each sample enterprise.

[0019] On the other hand, the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method for acquiring corporate behavior preferences described in any of the above embodiments when executing the program.

[0020] On the other hand, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method for obtaining corporate behavior preferences described in any of the above embodiments.

[0021] The method and device for obtaining corporate behavior preferences provided in the embodiments of the present invention integrate the N-degree implicit relationships that affect corporate behavior into graph features (centrality graph algorithm scores). Graph features are added as important features to the training of machine learning models. Compared with the current method of determining corporate behavior preferences only through the company's own information, the accuracy of determining corporate behavior preferences is greatly improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0023] Figure 1 It is a flowchart of the method for obtaining corporate behavior preferences provided in the first embodiment of the present invention.

[0024] Figure 2 It is a schematic diagram of an enterprise knowledge graph constructed according to the method for acquiring enterprise behavior preferences provided in the second embodiment of the present invention.

[0025] Figure 3 It is a partial flow chart of the method for obtaining corporate behavior preferences provided in the third embodiment of the present invention.

[0026] Figure 4 It is a flowchart of a method for acquiring corporate behavior preferences provided in the fourth embodiment of the present invention.

[0027] Figure 5 It is a partial flow chart of the method for obtaining corporate behavior preferences provided in the fifth embodiment of the present invention.

[0028] Figure 6 It is a partial flow chart of the method for obtaining corporate behavior preferences provided in the sixth embodiment of the present invention.

[0029] Figure 7 It is a flowchart of a method for acquiring corporate behavior preferences provided in the seventh embodiment of the present invention.

[0030] Figure 8 It is a schematic diagram of the operating position of each step in the method for obtaining corporate behavior preferences provided in the eighth embodiment of the present invention in the system.

[0031] Fig. 9 It is a schematic diagram of the structure of a device for acquiring corporate behavior preferences provided in the ninth embodiment of the present invention.

[0032] Fig.10It is a structural diagram of a device for acquiring corporate behavior preferences provided in the tenth embodiment of the present invention.

[0033] Fig.11 It is a structural diagram of a determination module in the device for acquiring corporate behavior preferences provided in the eleventh embodiment of the present invention.

[0034] Fig.12 It is a schematic diagram of the structure of a device for acquiring corporate behavior preferences provided in the twelfth embodiment of the present invention.

[0035] Fig.13 It is a structural diagram of a device for acquiring corporate behavior preferences provided in the thirteenth embodiment of the present invention.

[0036] Fig.14 It is a schematic diagram of the physical structure of an electronic device provided in the fourteenth embodiment of the present invention. DETAILED DESCRIPTION

[0037] In order to make the purpose, technical scheme and advantages of the embodiments of the present invention more clear, the embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings. Here, the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention. It should be noted that, in the absence of conflict, the embodiments in this application and the features in the embodiments can be combined with each other arbitrarily.

[0038] The execution subject of the method for obtaining corporate behavior preferences provided in the embodiment of the present invention includes but is not limited to a computer.

[0039] Figure 1 is a flow chart of a method for obtaining corporate behavior preferences provided by an embodiment of the present invention. Figure 1 As shown, the method for obtaining corporate behavior preferences provided by the embodiment of the present invention includes:

[0040] S101, constructing an enterprise knowledge graph based on the basic information and behavior information of each enterprise, wherein the enterprise knowledge graph has the enterprises as nodes and the relationships between the enterprises as edges;

[0041] In this step, the basic information of the enterprise may include information such as enterprise size, industry type, products held, business scope, registered assets, etc. The behavior information of the enterprise may include detailed transaction information of capital transfer in and out, industrial and commercial registration information, equity relationship information, investment information, credit financing and other information closely related to the behavior of the enterprise; the relationship between the enterprises may include a series of capital transaction relationships, shareholder relationships, investment relationships, etc.

[0042] Node and relationship information are extracted from the basic information and behavior information of each of the enterprises to construct the enterprise knowledge graph. The basic information and behavior information of each of the enterprises may be stored in different databases. Information from different sources can be integrated into the same database through an automated program, and nodes (i.e., each enterprise entity) and enterprise relationships can be extracted therefrom. The extracted node and relationship files can be stored in a hadoop cluster equipped with a JanusGraph graph database to construct an enterprise knowledge graph platform. Each of the nodes can be uniquely identified by the unified social credit code of the corresponding enterprise. The attributes of the node can include industry category, enterprise scale, registered assets, business scope, etc.; the relationship between each node can include a series of capital transaction relationships, shareholder relationships, investment relationships, etc., and the attributes of the relationship can include amount, shareholding ratio, etc. The start and end nodes contained in each relationship can be represented by the unique identifier of the corresponding enterprise.

[0043] Optionally, for users who need to conduct precision marketing for each of the above-mentioned enterprises, when constructing the enterprise knowledge graph to obtain the behavioral preferences of each enterprise, the above-mentioned enterprises can be divided into our company's customers and non-our company's customers, so that after determining the behavioral preferences of each of the above-mentioned enterprises, the corresponding marketing plan can be determined according to whether the enterprise is a customer of our company. For example, when the user who needs to conduct precision marketing for each of the above-mentioned enterprises is a banking system, the attributes of each node in the enterprise knowledge graph may include the bank of account opening and the customer number of the bank of account opening. The enterprise knowledge graph constructed accordingly can be as follows: Figure 2 shown.

[0044] S102, calculating the centrality graph algorithm score of each enterprise under each relationship type according to the relationship type between the enterprises in the enterprise knowledge graph, wherein there is at least one relationship between two enterprises linked in the enterprise knowledge graph;

[0045] In this step, the centrality graph algorithm score of each enterprise is calculated according to the dimension of relationship type. Specifically, for each enterprise under each relationship type in the enterprise knowledge graph, based on the centrality graph algorithm in the knowledge graph (the core of this algorithm considers the number and quality of the incoming relationships between nodes, thereby determining the importance of the nodes), the centrality graph algorithm score of each enterprise under the relationship type is calculated to determine the importance of each enterprise under the relationship type.

[0046] S103. Input the basic information, behavior information and centrality graph algorithm scores of each enterprise into a pre-trained behavior preference prediction model to obtain the behavior preference of the enterprise, wherein the behavior preference prediction model is obtained after training based on the basic information, behavior information, centrality graph algorithm scores of the sample enterprise and the behavior preference labels predetermined by the sample enterprise.

[0047] In this step, after the behavioral preference model is obtained through training based on the basic information, behavioral information, centrality graph algorithm scores of the sample enterprises and the behavioral preference labels predetermined by the sample enterprises, the basic information, behavioral information and centrality graph algorithm scores of an enterprise are directly input into the behavioral preference prediction model to obtain the behavioral preference of the enterprise.

[0048] The above information of the sample enterprises is used as sample data for training the behavior preference prediction model; the sample enterprises may be other enterprises different from the enterprises in the above step S101, or may include at least a part of the enterprises in the above step S101. The types of behavior preference labels of the sample enterprises may include fund collection type, equity investment type, industrial operation type, etc., and the behavior preference labels of the sample enterprises may be manually set.

[0049] The method for obtaining corporate behavior preferences provided in an embodiment of the present invention integrates the N-degree implicit relationships that affect corporate behavior into graph features (centrality graph algorithm scores). Graph features are added as important features to the training of machine learning models. Compared with the current method of determining corporate behavior preferences only through the company's own information, the accuracy of determining corporate behavior preferences is greatly improved.

[0050] like Figure 3 As shown, optionally, in the above embodiment, the behavior preference prediction model obtained by training based on the basic information, behavior information, centrality graph algorithm scores of the sample enterprises and the behavior preference labels predetermined by the sample enterprises may include:

[0051] S001. Construct a sample enterprise knowledge graph based on the basic information and behavior information of each sample enterprise, wherein the sample enterprise knowledge graph has each sample enterprise as a node and the relationship between each sample enterprise as an edge;

[0052] The execution process of this step is similar to the execution process of the above step S101, and will not be repeated here.

[0053] S002. Calculate the centrality graph algorithm score of each sample enterprise under each relationship type according to the relationship type between the sample enterprises in the sample enterprise knowledge graph, wherein there is at least one relationship between two sample enterprises linked in the sample enterprise knowledge graph;

[0054] The execution process of this step is similar to the execution process of the above step S102, and will not be repeated here.

[0055] S003. According to the basic information, behavior information, centrality graph algorithm scores and predetermined behavior preference labels of each of the sample enterprises, a behavior preference prediction model is trained.

[0056] In this embodiment, when training the enterprise preference model, the input features "basic information, behavioral information and centrality algorithm scores of sample enterprises" include numerical features and nominal features. The numerical features can be normalized to obtain numerical feature vectors (such as registered capital data, etc.), and the nominal features such as industry category, enterprise type, etc. can be One-Hot encoded to obtain nominal feature vectors. The numerical feature vector and the nominal feature vector are combined to form an input feature vector.

[0057] The sample size of the data set of the enterprise preference model can be 10,000 (that is, the number of sample enterprises can be 10,000), and the training set, validation set and test set are set in a ratio of 7:2:1. The XGBoost algorithm (with the base learner being CART) can be used to learn and train the input features composed of enterprise category, industry type, registered capital, enterprise operation centrality graph algorithm score, funds allocation centrality graph algorithm score, equity relationship centrality graph algorithm score, investment centrality graph algorithm score, and industrial chain centrality graph algorithm score to predict the enterprise's behavioral preference type (fund collection type, equity investment type, industrial operation type, etc.).

[0058] A feedback mechanism can also be provided for the prediction accuracy of the behavior preference prediction model. For example, a graph query module based on the system application layer provides a feedback mechanism to feed back whether the behavior preferences of sample enterprises are predicted correctly to the behavior preference prediction model construction module, and add sample sets to continuously iterate and optimize the model.

[0059] Optionally, in the above embodiment, the calculating the centrality graph algorithm score of each enterprise under each relationship type according to the relationship type between the enterprises in the enterprise knowledge graph may include:

[0060] For each enterprise under each relationship type in the enterprise knowledge graph, the centrality graph algorithm score of each enterprise is calculated using the following centrality graph algorithm score calculation formula (PageRank algorithm):

[0061]

[0062] Among them, T1~T n are neighbor nodes that have a link relationship with node u, PR(u), PR(T1)...PR(T n ) represent nodes u, T1...T n The centrality graph algorithm score, C(T n ) is the node T n The degree refers to the number of neighboring nodes that the node is linked to; d is the damping coefficient, and (1-d) represents the probability that node u is linked to.

[0063] In this embodiment, the PageRank algorithm: by randomly traversing the enterprise knowledge graph, the score of a node is propagated to its neighboring nodes, and the frequency of visiting each node in these traversal processes is calculated (considering the number of neighboring nodes and the attributes of the link relationship). For all nodes, the frequency value of all surrounding nodes linking to it can be calculated, which is used as the score of the PageRank algorithm (centrality graph algorithm score).

[0064] The damping coefficient d may be set to 0.85, and for nodes without link-in relationships, the initial PR value is 0.15. In the process of applying the above formula to calculate the centrality graph algorithm score of node u, the PR score of each node is continuously updated until convergence or the number of calculations is reached.

[0065] If the score of node u is high, it means that this node is more central and more important. By calculating the centrality graph algorithm score of enterprises under different relationship types, we can obtain the enterprise's capital transaction (enterprise operation, industrial chain, etc.) graph characteristics (capital transaction centrality graph algorithm score), equity relationship graph characteristics (equity relationship centrality graph algorithm score), investment relationship graph characteristics (investment relationship centrality graph algorithm score), credit relationship graph characteristics (credit relationship centrality graph algorithm score), etc.

[0066] like Figure 4 As shown, optionally, after inputting the basic information, behavior information and centrality graph algorithm scores of each enterprise into a pre-trained behavior preference prediction model to obtain the behavior preference of the enterprise, the method may further include:

[0067] S104. According to the behavior preference type of the enterprise, determine the marketing path for the enterprise based on the knowledge graph shortest path algorithm.

[0068] In this embodiment, according to the behavior preference category of the enterprise, the marketing path for the enterprise is determined based on the knowledge graph shortest path algorithm, and the marketing path is determined for the specific relationship category. Specifically, for different preference types of the enterprise, the corresponding related relationship path is found. For example, for a banking system that needs to conduct precision marketing for the enterprise, if the enterprise is an enterprise that our bank needs to retain customers, the shortest path determined by its behavior preference is used to determine the marketing community and path for targeted marketing. If the enterprise is an enterprise that our bank needs to expand its customers, according to its behavior preference, the association between the enterprise and our customers is found through the shortest path, and customers are expanded with the help of our customer resources.

[0069] like Figure 5As shown, optionally, in the above embodiment, determining the marketing path for the enterprise based on the knowledge graph shortest path algorithm according to the behavior preference type of the enterprise may specifically include:

[0070] S1041. Obtaining enterprise relationship types related to the behavior preference types of the enterprise in the enterprise knowledge graph;

[0071] In this step, the corporate behavior preferences predicted by the behavior preference prediction model are mapped to the relationship types of the corporate knowledge graph, such as investment relationships for investment-type companies, corporate operating relationships for business operations, and shareholder relationships for shareholder types.

[0072] S1042. Calculate the shortest path between the enterprise and other enterprises under the relationship type in the enterprise knowledge graph based on the knowledge graph shortest path algorithm;

[0073] In this step, based on the graph query in the knowledge graph, the relationship between enterprises can be restricted to the relationship type determined in step S1041, and the single-source shortest path from the enterprise to other enterprises can be found; specifically, the shortest path from the enterprise to other enterprises can be queried using depth-first traversal, that is, starting from the starting node and taking other nodes in the graph as the end nodes, and calling the shortest path one by one to achieve it.

[0074] S1043. Determine a marketing path for the enterprise based on the shortest path between other enterprises and the enterprise.

[0075] In this step, for users who need to conduct precision marketing for each enterprise in the enterprise knowledge graph, some of the enterprises may already be customers of the user, while others are not yet customers of the user. In order to facilitate the user to determine the marketing path for the enterprise according to the shortest path, the relationship between each enterprise and the user can be added to the node attribute corresponding to the enterprise when constructing the enterprise knowledge graph; for example, when the user is a bank system, the bank where each enterprise opens an account and the code relative to the bank where the account opens can be added to the attribute of the corresponding node. In this way, when determining the marketing path for the enterprise according to the shortest path in this step, the enterprises that have a relevant relationship with our bank in each shortest path can be used as entry points to conduct precision marketing for the enterprise as the starting node, thereby realizing the association with the optimal marketing path through the enterprise behavior preference characteristics, and solving the problem of unclear marketing path and low success rate for enterprise customers in the prior art.

[0076] like Figure 6 As shown, optionally, in the above embodiment, determining the marketing path for the enterprise according to the shortest path between each other enterprise and the enterprise may include:

[0077] S10431, obtaining the path depth of each of the shortest paths and the centrality graph score of the key nodes in each of the shortest paths;

[0078] S10432. The shortest path in which the path depth is less than the target depth and the centrality graph score of the key node in the path is greater than the target score is used as the marketing path for the enterprise.

[0079] In this embodiment, the path depth of each shortest path refers to the number of links required from the terminating node to the starting node in the shortest path; taking into account the feasibility of the marketing path, corresponding target depths (for example, controlling the path depth to be no more than 3) can be set for different business scenarios of customer acquisition and customer retention, and the target scores of the key nodes in the shortest path (which may be user enterprises that need to conduct precision marketing for the enterprises corresponding to the starting nodes in the shortest path) (for example, controlling the centrality graph algorithm score to be no less than 0.4) are used to screen the shortest paths to obtain the marketing path.

[0080] like Figure 7 As shown, optionally, after determining the marketing path for the enterprise based on the knowledge graph shortest path algorithm according to the behavior preference type of the enterprise, the method further includes:

[0081] S105. Display the marketing path in the form of a visual enterprise knowledge graph.

[0082] In this embodiment, the optimal marketing path associated with the enterprise behavior preference characteristics is displayed in the form of a graph, presenting a detailed and clear marketing path so that the marketing behavior can be traced.

[0083] like Figure 8 As shown, in the above embodiments, the acquisition and preprocessing process of the basic information and behavioral information of each of the enterprises can be run in the data collection layer of the system, the construction of the enterprise knowledge graph and the training process of the behavioral preference prediction model can be run in the model construction layer of the system, and the graph query and marketing path display process can be run in the application layer of the system.

[0084] Fig. 9 is a schematic diagram of the structure of a device for acquiring corporate behavior preferences provided by an embodiment of the present invention. Fig. 9As shown, the device for acquiring enterprise behavior preference provided by the embodiment of the present invention includes: a graph construction module 21, which is used to construct an enterprise knowledge graph based on the basic information and behavior information of each enterprise, wherein the enterprise knowledge graph has each enterprise as a node and the relationship between each enterprise as an edge; a calculation module 22, which is used to calculate the centrality graph algorithm score of each enterprise under each relationship type according to the relationship type between the enterprises in the enterprise knowledge graph, wherein there is at least one relationship between two linked enterprises in the enterprise knowledge graph; an acquisition module 23, which is used to input the basic information, behavior information and each centrality graph algorithm score of each enterprise into a pre-trained behavior preference prediction model to obtain the behavior preference of the enterprise, wherein the behavior preference prediction model is obtained after training based on the basic information, behavior information, each centrality graph algorithm score of the sample enterprise and the behavior preference labels predetermined by the sample enterprise.

[0085] Optionally, the calculation module is specifically used to: for each enterprise under each relationship type in the enterprise knowledge graph, calculate the centrality graph algorithm score of each enterprise using the following centrality graph algorithm score calculation formula: Among them, T1~T n are neighbor nodes that have a link relationship with node u, PR(u), PR(T1)...PR(T n ) represent nodes u, T1...T n The centrality graph algorithm score, C(T n ) is the node T n The degree refers to the number of neighboring nodes that the node is linked to; d is the damping coefficient, and (1-d) represents the probability that node u is linked to.

[0086] like Fig.10 As shown, optionally, the device also includes: a determination module 24, which is used to determine a marketing path for the enterprise based on the knowledge graph shortest path algorithm according to the behavior preference type of the enterprise.

[0087] like Fig.11 As shown, optionally, the determination module specifically includes: an acquisition unit 241, used to obtain the enterprise relationship type related to the behavior preference type of the enterprise in the enterprise knowledge graph; a calculation unit 242, used to calculate the shortest path between each other enterprise under the relationship type in the enterprise knowledge graph and the enterprise based on the knowledge graph shortest path algorithm; a determination unit 243, used to determine the marketing path for the enterprise based on the shortest path between each other enterprise and the enterprise.

[0088] Optionally, the determination unit is specifically used to: obtain the path depth of each of the shortest paths, and the centrality graph score of the key nodes in each of the shortest paths; and use the shortest path in which the path depth is less than the target depth and the centrality graph score of the key nodes in the path is greater than the target score as the marketing path for the enterprise.

[0089] like Fig.12 As shown, optionally, the device also includes: a display module 25, which is used to display the marketing path in the form of a visual enterprise knowledge graph.

[0090] like Fig.13 As shown, optionally, the device also includes a model building module 26, and the model building module 26 is used to: build a sample enterprise knowledge graph based on the basic information and behavior information of each sample enterprise, wherein the sample enterprise knowledge graph has each sample enterprise as a node and the relationship between each sample enterprise as an edge; calculate the centrality graph algorithm score of each sample enterprise under each relationship type based on the relationship type between each sample enterprise in the sample enterprise knowledge graph, wherein there is at least one relationship between two sample enterprises linked in the sample enterprise knowledge graph; and train a behavior preference prediction model based on the basic information, behavior information, each centrality graph algorithm score and predetermined behavior preference labels of each sample enterprise.

[0091] The embodiments of the apparatus provided in the embodiments of the present invention can be specifically used to execute the processing flow of the above-mentioned method embodiments, and its functions are not described in detail here, and reference can be made to the detailed description of the above-mentioned method embodiments.

[0092] It should be noted that the method and device for obtaining corporate behavior preferences provided in the embodiments of the present invention can be used in the financial field, and can also be used in any technical field other than the financial field. The embodiments of the present invention do not limit the application field of the method and device for obtaining corporate behavior preferences.

[0093] Fig.14 A schematic diagram of the physical structure of an electronic device provided by an embodiment of the present invention is shown in FIG. Fig.14As shown, the electronic device may include: a processor 301, a communication interface 302, a memory 303 and a communication bus 304, wherein the processor 301, the communication interface 302 and the memory 303 complete mutual communication through the communication bus 304. The processor 301 may call the logic instructions in the memory 303 to execute the following method: construct an enterprise knowledge graph according to the basic information and behavior information of each enterprise, wherein the enterprise knowledge graph has each enterprise as a node and the relationship between each enterprise as an edge; calculate the centrality graph algorithm score of each enterprise under each relationship type according to the relationship type between each enterprise in the enterprise knowledge graph, wherein at least one relationship exists between two linked enterprises in the enterprise knowledge graph; input the basic information, behavior information and each centrality graph algorithm score of each enterprise into a pre-trained behavior preference prediction model to obtain the behavior preference of the enterprise, wherein the behavior preference prediction model is obtained after training based on the basic information, behavior information, each centrality graph algorithm score of the sample enterprise and the behavior preference label pre-determined by the sample enterprise.

[0094] In addition, the logic instructions in the above-mentioned memory 303 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0095] The present embodiment discloses a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the methods provided by the above-mentioned method embodiments, for example, including: constructing an enterprise knowledge graph based on basic information and behavior information of each enterprise, wherein the enterprise knowledge graph has each enterprise as a node and the relationship between each enterprise as an edge; calculating the centrality graph algorithm score of each enterprise under each relationship type based on the relationship type between the enterprises in the enterprise knowledge graph, wherein there is at least one relationship between two linked enterprises in the enterprise knowledge graph; inputting the basic information, behavior information and each centrality graph algorithm score of each enterprise into a pre-trained behavior preference prediction model to obtain the behavior preference of the enterprise, wherein the behavior preference prediction model is obtained after training based on the basic information, behavior information, each centrality graph algorithm score of the sample enterprise and the behavior preference labels predetermined by the sample enterprise.

[0096] The present embodiment provides a computer-readable storage medium, which stores a computer program, and the computer program enables the computer to execute the methods provided by the above-mentioned method embodiments, for example, including: constructing an enterprise knowledge graph based on the basic information and behavior information of each enterprise, wherein the enterprise knowledge graph has each enterprise as a node and the relationship between each enterprise as an edge; calculating the centrality graph algorithm score of each enterprise under each relationship type based on the relationship type between the enterprises in the enterprise knowledge graph, wherein there is at least one relationship between two linked enterprises in the enterprise knowledge graph; inputting the basic information, behavior information and each centrality graph algorithm score of each enterprise into a pre-trained behavior preference prediction model to obtain the behavior preference of the enterprise, wherein the behavior preference prediction model is obtained after training based on the basic information, behavior information, each centrality graph algorithm score of the sample enterprise and the behavior preference labels predetermined by the sample enterprise.

[0097] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0098] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0099] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0100] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0101] In the description of this specification, the description with reference to the terms "one embodiment", "a specific embodiment", "some embodiments", "for example", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0102] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for obtaining corporate behavior preferences, characterized in that: include: According to the basic information and behavior information of each enterprise, an enterprise knowledge graph is constructed, wherein the enterprise knowledge graph has the enterprises as nodes and the relationships between the enterprises as edges; According to the relationship type between the enterprises in the enterprise knowledge graph, the centrality graph algorithm score of each enterprise under each relationship type is calculated, wherein there is at least one relationship between two enterprises linked in the enterprise knowledge graph, and the relationship type includes at least one of the following: capital transaction relationship, equity relationship, investment relationship, and credit relationship; Inputting the basic information, behavior information and centrality graph algorithm scores of each enterprise into a pre-trained behavior preference prediction model to obtain the behavior preference of the enterprise, wherein the behavior preference prediction model is obtained after training based on the basic information, behavior information, centrality graph algorithm scores of the sample enterprise and the behavior preference labels predetermined by the sample enterprise; The calculating of the centrality graph algorithm score of each enterprise under each relationship type according to the relationship type between the enterprises in the enterprise knowledge graph comprises: For each enterprise under each relationship type in the enterprise knowledge graph, the centrality graph algorithm score of each enterprise is calculated using the following centrality graph algorithm score calculation formula: Among them, T1~T n are neighbor nodes that have a link relationship with node u, PR(u), PR(T1)...PR(T n ) represent nodes u, T1...T n The centrality graph algorithm score, C(T n ) is the node T n The degree refers to the number of neighboring nodes that the node is linked to; d is the damping coefficient, and (1-d) represents the probability that node u is linked to.

2. The method according to claim 1, characterized in that After inputting the basic information, behavior information and centrality graph algorithm scores of each enterprise into a pre-trained behavior preference prediction model to obtain the behavior preference of the enterprise, the method further includes: According to the behavioral preference type of the enterprise, a marketing path for the enterprise is determined based on the knowledge graph shortest path algorithm.

3. The method according to claim 2, characterized in that Determining the marketing path for the enterprise based on the knowledge graph shortest path algorithm according to the behavior preference type of the enterprise includes: Obtaining enterprise relationship types related to the behavior preference types of the enterprise in the enterprise knowledge graph; Based on the knowledge graph shortest path algorithm, calculate the shortest path between the enterprise and other enterprises under the relationship type in the enterprise knowledge graph; A marketing path for the enterprise is determined based on the shortest path between other enterprises and the enterprise.

4. The method according to claim 3, characterized in that Determining the marketing path for the enterprise according to the shortest path between each other enterprise and the enterprise includes: Obtaining the path depth of each of the shortest paths and the centrality graph scores of the key nodes in each of the shortest paths; The shortest path in which the path depth is less than the target depth and the centrality graph score of the key node in the path is greater than the target score is used as the marketing path for the enterprise.

5. The method according to claim 2, characterized in that: After determining the marketing path for the enterprise based on the knowledge graph shortest path algorithm according to the behavior preference type of the enterprise, the method further includes: The marketing path is presented in the form of a visual enterprise knowledge graph.

6. The method according to claim 1, characterized in that The behavior preference prediction model is obtained by training based on the basic information, behavior information, algorithm scores of each centrality graph and the behavior preference labels predetermined by the sample enterprises, including: Construct a sample enterprise knowledge graph based on the basic information and behavior information of each sample enterprise, wherein the sample enterprise knowledge graph has each sample enterprise as a node and the relationship between each sample enterprise as an edge; According to the relationship type between the sample enterprises in the sample enterprise knowledge graph, the centrality graph algorithm score of each sample enterprise under each relationship type is calculated, wherein there is at least one relationship between two sample enterprises linked in the sample enterprise knowledge graph; A behavior preference prediction model is trained based on the basic information, behavior information, centrality graph algorithm scores and predetermined behavior preference labels of each of the sample enterprises.

7. A device for acquiring corporate behavior preferences, characterized in that: include: A graph construction module is used to construct an enterprise knowledge graph based on the basic information and behavior information of each enterprise, wherein the enterprise knowledge graph has each enterprise as a node and the relationship between each enterprise as an edge; A calculation module, for calculating the centrality graph algorithm score of each enterprise under each relationship type according to the relationship type between the enterprises in the enterprise knowledge graph, wherein there is at least one relationship between two enterprises linked in the enterprise knowledge graph, and the relationship type includes at least one of the following: capital transaction relationship, equity relationship, investment relationship, and credit relationship; An acquisition module is used to input the basic information, behavior information and centrality graph algorithm scores of each enterprise into a pre-trained behavior preference prediction model to obtain the behavior preference of the enterprise, wherein the behavior preference prediction model is obtained after training based on the basic information, behavior information, centrality graph algorithm scores of the sample enterprise and the behavior preference labels predetermined by the sample enterprise; The calculation module is specifically used for: For each enterprise under each relationship type in the enterprise knowledge graph, the centrality graph algorithm score of each enterprise is calculated using the following centrality graph algorithm score calculation formula: Among them, T1~T n are neighbor nodes that have a link relationship with node u, PR(u), PR(T1)...PR(T n ) represent nodes u, T1...T n The centrality graph algorithm score, C(T n ) is the node T n The degree refers to the number of neighboring nodes that the node is linked to; d is the damping coefficient, and (1-d) represents the probability that node u is linked to.

8. The device according to claim 7, characterized in that The device also includes: A determination module is used to determine a marketing path for the enterprise based on the knowledge graph shortest path algorithm according to the behavioral preference type of the enterprise.

9. The device according to claim 8, characterized in that The determination module specifically includes: An acquisition unit, used to acquire an enterprise relationship type related to the behavior preference type of the enterprise in the enterprise knowledge graph; A calculation unit, used for calculating the shortest path between the enterprise and each other enterprise under the relationship type in the enterprise knowledge graph based on the knowledge graph shortest path algorithm; The determination unit is used to determine a marketing path for the enterprise based on the shortest path between other enterprises and the enterprise.

10. The device according to claim 9, characterized in that The determining unit is specifically used for: Obtaining the path depth of each of the shortest paths and the centrality graph scores of the key nodes in each of the shortest paths; The shortest path in which the path depth is less than the target depth and the centrality graph score of the key node in the path is greater than the target score is used as the marketing path for the enterprise.

11. The device according to claim 8, characterized in that The device also includes: The display module is used to display the marketing path in the form of a visual enterprise knowledge graph.

12. The device according to claim 7, characterized in that The device also includes a model building module, and the model building module is used to: Construct a sample enterprise knowledge graph based on the basic information and behavior information of each sample enterprise, wherein the sample enterprise knowledge graph has each sample enterprise as a node and the relationship between each sample enterprise as an edge; According to the relationship type between the sample enterprises in the sample enterprise knowledge graph, the centrality graph algorithm score of each sample enterprise under each relationship type is calculated, wherein there is at least one relationship between two sample enterprises linked in the sample enterprise knowledge graph; A behavior preference prediction model is trained based on the basic information, behavior information, centrality graph algorithm scores and predetermined behavior preference labels of each of the sample enterprises.

13. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

14. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images