Corporate credit assessment methods, devices, electronic equipment and readable storage media

Abstract

This application provides a method, apparatus, electronic device, and readable storage medium for enterprise credit assessment, belonging to the field of enterprise credit assessment technology. The method includes: acquiring multi-dimensional supply chain data; preprocessing the multi-dimensional supply chain data to obtain multi-dimensional related data sources; constructing a payment chain knowledge graph based on the multi-dimensional related data sources, and obtaining node indicators of the payment chain knowledge graph; and calculating an enterprise credit score based on the node indicators and key payment data using a pre-constructed final integrated credit assessment model. Thus, by calculating the enterprise credit score using the pre-constructed final integrated credit assessment model based on the node indicators and key payment data of the payment chain knowledge graph, the accuracy of enterprise credit prediction and the model's generalization ability are improved.

Description

Technical Field

[0001] This application relates to the field of corporate credit assessment technology, and in particular to a corporate credit assessment method, apparatus, electronic device and readable storage medium. Background Technology

[0002] Currently, with rising consumer demand and increasingly fierce market competition, supply chains are evolving towards globalization and greater complexity. Enterprises increasingly rely on smarter and more efficient supply chains for development. The fintech industry offers a supply chain finance model where core enterprises leverage their upstream and downstream resources to fully integrate supply chain and customer resources, providing financing channels for all participants in the supply chain. The essence of supply chain finance is credit-driven financial innovation, making illiquid assets tradable. Supply chain finance can be used to pay for new loans, thereby increasing lending to SMEs.

[0003] In production and operation, organized and regular payment behaviors continuously exist between enterprises. The transaction chain relationships between enterprises, depicted by payment and receipt data and their flow, form a vast payment chain graph system, containing enormous financial service value. In the intricate transaction behavior of enterprises, the payment chain is the core standard for reflecting and evaluating the health and quality of the supply chain between enterprises. It provides the most valuable data support and basis for the generation and transmission of credit chains, and also offers broad prospects for the application of blockchain.

[0004] With my country's rapid economic development and the increasing openness of its financial markets, credit assessment of core enterprises within the supply chain is becoming increasingly important. The fintech industry provides enterprise credit big data assessment solutions. These solutions involve professional credit rating agencies or algorithm engineers using expert judgment or technical analysis methods to comprehensively analyze and evaluate an enterprise's ability to fulfill various commitments and its creditworthiness, expressing this information in simple and clear symbols or text to meet societal market needs. However, existing enterprise credit assessment models suffer from low accuracy in predicting creditworthiness and insufficient model generalization ability. Summary of the Invention

[0005] To address the aforementioned technical problems, embodiments of this application provide a method, apparatus, electronic device, and readable storage medium for enterprise credit assessment.

[0006] In a first aspect, embodiments of this application provide a method for enterprise credit assessment, the method comprising:

[0007] Obtain multi-dimensional data from the supply chain;

[0008] The multidimensional supply chain data is preprocessed to obtain a multidimensional related data source;

[0009] A payment chain knowledge graph is constructed based on the multi-dimensional associated data sources, and node indicators of the payment chain knowledge graph are obtained.

[0010] The enterprise credit score is calculated based on the node indicators and key payment data through a pre-built final integrated credit assessment model.

[0011] In one embodiment, acquiring multidimensional supply chain data includes:

[0012] Obtain multiple basic data tables of the supply chain, and obtain multidimensional correlation factors for each basic data table; obtain derived data for each basic data table based on the multidimensional correlation factors;

[0013] The supply chain multidimensional data is composed of multiple basic data tables and multiple derived data tables. In one embodiment, the preprocessing of the supply chain multidimensional data includes:

[0014] The multidimensional data of the supply chain is preprocessed using a synthetic minority class oversampling algorithm.

[0015] In one embodiment, constructing a payment chain knowledge graph based on the multi-dimensional associated data sources includes:

[0016] Multiple knowledge units are acquired, and the unit relationships between each knowledge unit are obtained based on the multi-dimensional associated data source.

[0017] The payment chain knowledge graph is constructed based on the unit relationships between various knowledge units.

[0018] In one embodiment, obtaining the node metrics of the payment chain knowledge graph includes:

[0019] Obtain the sum of the distances between each node and other nodes in the payment chain knowledge graph;

[0020] The importance of each node is determined based on its distance from other nodes;

[0021] Obtain the sum of the shortest paths or the average shortest distance between each node and other nodes in the payment chain knowledge graph;

[0022] The node density is determined based on the sum of the shortest paths or the average shortest distance.

[0023] In one embodiment, constructing the credit assessment model includes:

[0024] An initial integrated credit assessment model is constructed based on multiple initial scenario sub-modules and an initial integration sub-module.

[0025] The supply chain sample data is divided into a training set and a test set;

[0026] The initial integrated credit assessment model is trained using the training set to adjust its parameters, thereby obtaining a modified integrated credit assessment model.

[0027] The modified integrated credit assessment model was tested using the test set, and the test results were obtained.

[0028] The parameters of the modified integrated credit assessment model are adjusted based on the test results to obtain the final integrated credit assessment model.

[0029] In one embodiment, the method further includes:

[0030] Enterprise tag topics are generated based on the payment chain knowledge graph.

[0031] Secondly, embodiments of this application provide a corporate credit assessment device, the device comprising:

[0032] The acquisition module is used to acquire multi-dimensional data from the supply chain.

[0033] The processing module is used to preprocess the multidimensional supply chain data to obtain multidimensional related data sources;

[0034] The construction module is used to construct a payment chain knowledge graph based on the multi-dimensional associated data sources and obtain the node indicators of the payment chain knowledge graph.

[0035] The calculation module is used to calculate the enterprise credit score based on the node indicators and key payment data through a pre-built final integrated credit assessment model.

[0036] Thirdly, embodiments of this application provide an electronic device, including a memory and a processor, wherein the memory is used to store a computer program, and the computer program executes the enterprise credit assessment method provided in the first aspect when the processor is running.

[0037] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when run on a processor, executes the enterprise credit assessment method provided in the first aspect.

[0038] The enterprise credit assessment method, apparatus, electronic device, and readable storage medium provided in this application acquire multi-dimensional supply chain data; preprocess the multi-dimensional supply chain data to obtain multi-dimensional related data sources; construct a payment chain knowledge graph based on the multi-dimensional related data sources, and obtain node indicators of the payment chain knowledge graph; calculate an enterprise credit score based on the node indicators and key payment data using a pre-constructed final integrated credit assessment model. In this way, by calculating an enterprise credit score based on the node indicators and key payment data of the payment chain knowledge graph using a pre-constructed final integrated credit assessment model, the accuracy of enterprise credit prediction is improved, and the model's generalization ability is enhanced. Attached Figure Description

[0039] To more clearly illustrate the technical solutions of this application, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of this application and should not be considered as a limitation on the scope of protection of this application. In the various drawings, similar components are numbered similarly.

[0040] Figure 1 One of the flowcharts of the enterprise credit assessment method provided in this application is shown;

[0041] Figure 2 This is a second schematic flowchart of the enterprise credit assessment method provided in the embodiments of this application;

[0042] Figure 3 The third schematic diagram of the enterprise credit assessment method provided in the embodiments of this application is shown;

[0043] Figure 4 This paper shows one of the structural schematic diagrams of the enterprise credit assessment device provided in the embodiments of this application;

[0044] Figure 5 One of the structural schematic diagrams of the electronic device provided in the embodiments of this application is shown.

[0045] Icons: 400 - Enterprise Credit Assessment Device, 401 - Acquisition Module, 402 - Processing Module, 403 - Construction Module, 404 - Calculation Module;

[0046] 500 - Electronic equipment, 501 - Transceiver, 502 - Processor, 503 - Memory. Detailed Implementation

[0047] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0048] The components of the embodiments of this application described and illustrated in the accompanying drawings can be arranged and designed in a variety of different configurations. Therefore, the following detailed description of the embodiments of this application provided in the drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0049] In the following, the terms “comprising,” “having,” and their cognates, which may be used in various embodiments of this application, are intended only to indicate a particular feature, number, step, operation, element, component, or combination thereof, and should not be construed as excluding, firstly, the presence of one or more other features, numbers, steps, operations, elements, components, or combinations thereof, or adding the possibility of one or more features, numbers, steps, operations, elements, components, or combinations thereof.

[0050] Furthermore, the terms "first," "second," and "third" are used only to distinguish descriptions and should not be interpreted as indicating or implying relative importance.

[0051] Unless otherwise specified, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which the various embodiments of this application pertain. Terms (such as those defined in commonly used dictionaries) shall be interpreted as having the same meaning as in their contextual meaning in the relevant technical field and shall not be construed as having an idealized or overly formal meaning, unless clearly defined in the various embodiments of this application.

[0052] Example 1

[0053] This application provides a method for corporate credit assessment.

[0054] See Figure 1 The corporate credit assessment method includes steps S101 to S104, which will be described in detail below.

[0055] Step S101: Obtain multi-dimensional supply chain data.

[0056] In the fintech industry, multi-dimensional supply chain data can be obtained from existing customer data within a bank's supply chain. This multi-dimensional supply chain data can verify the authenticity of transactions between upstream and downstream enterprises. Based on this multi-dimensional supply chain data, a data support system for enterprise payment capabilities and credit can be built to provide payment and financing services to enterprises. It can also assist banks in enterprise credit rating, default prediction, and credit line extension.

[0057] See Figure 2 Step S101 includes:

[0058] Step S1011: Obtain multiple basic data tables for the supply chain, and obtain the multi-dimensional data from each basic data table.

[0059] Related factors;

[0060] Step S1012: Obtain the derived data of each of the basic data tables based on the multidimensional correlation factors;

[0061] Step S1013: Combine the multiple basic data tables and the multiple derived data to form the supply chain multidimensional data.

[0062] As an example, multiple basic data tables can be understood as multiple types of basic data tables, such as intra-bank settlement data, intra-bank savings data, inter-bank credit data, order data, invoice data, payment data, etc.

[0063] In this embodiment, multiple basic data tables and multiple derived data are integrated to obtain multidimensional supply chain data. The integration process mainly involves associating various types of tables to obtain multidimensional supply chain data from the basic data tables and their derived associations.

[0064] For example, order data can be examined sequentially for the total order amount of the previous n months, where n = [0, 1, 2, 3, 6, 12], and n = i represents the current month, where i is a value of 0, 1, 2, 3, 6, or 12. Derived data can include indicators such as the average order amount per order in the previous n months, the total number of orders in the previous n months, the order frequency in the previous n months, the total number of orders in the previous n months, and the time interval between the most recent order. Similarly, other derived dimensions of data can be combined.

[0065] Step S102: Preprocess the multidimensional data of the supply chain to obtain a multidimensional related data source.

[0066] In one embodiment, step S102 includes:

[0067] The multidimensional data of the supply chain is preprocessed using a synthetic minority class oversampling algorithm.

[0068] In this embodiment, the Synthetic Minority Oversampling Technique (SMOTE) is used to process imbalanced data. The SMOTE imbalanced data processing steps are as follows: For each sample x in the minority class, the distance from each sample to all samples in the minority class sample set is calculated using Euclidean distance to obtain its k-nearest neighbors. For example, let k=5. A sampling ratio is set according to the imbalance ratio to determine the sampling multiplier N. For each minority class sample x, several samples are randomly selected from its k-nearest neighbors. Assume the selected nearest neighbors are... For each randomly selected nearest neighbor New samples are constructed from the original samples using the following formula.

[0069] ;

[0070] The SMOTE algorithm mainly focuses on sample construction. Generally, the proportion of negative samples is much smaller than that of positive samples. By generating more negative samples through the SMOTE algorithm, the ratio of positive to negative samples can be balanced.

[0071] Step S103: Construct a payment chain knowledge graph based on the multi-dimensional associated data source and obtain the node indicators of the payment chain knowledge graph.

[0072] In one embodiment, the step S103 of constructing a payment chain knowledge graph based on the multi-dimensional associated data source includes:

[0073] Step S1031: Obtain multiple knowledge units and obtain the unit relationships between each knowledge unit based on the multi-dimensional associated data source;

[0074] Step S1032: Construct the payment chain knowledge graph based on the unit relationships between the various knowledge units.

[0075] In this embodiment, NLP techniques such as sentence segmentation, part-of-speech tagging, and entity recognition are used to achieve entity recognition, relation extraction, and knowledge graph construction.

[0076] As an example, the construction of a payment chain knowledge graph may include: data integration from multi-dimensional related data sources; preprocessing of multi-dimensional related data sources, including word segmentation, stop word removal, deduplication, and errata correction; selection of knowledge units based on core enterprises, distributors, and sub-suppliers; construction of unit relationships for knowledge units, including payment relationships, guarantee relationships, equity relationships, and supply chain relationships; standardization of unit relationships; semantic relationship analysis of the standardized unit relationships through cluster analysis and latent semantic analysis; and visualization of the semantic relationship analysis results to obtain a visualized payment chain knowledge graph.

[0077] In this embodiment, the payment chain knowledge graph can provide visualization and retrieval functions. Combining the payment chain graph with the relationships between neighboring nodes can enrich the feature expression of nodes and expand their dimensions.

[0078] In one embodiment, obtaining the node indicators of the payment chain knowledge graph in step S103 includes:

[0079] Obtain the sum of the distances between each node and other nodes in the payment chain knowledge graph;

[0080] The importance of each node is determined based on its distance from other nodes;

[0081] Obtain the sum of the shortest paths or the average shortest distance between each node and other nodes in the payment chain knowledge graph;

[0082] The node density is determined based on the sum of the shortest paths or the average shortest distance.

[0083] The exemplary topology based on the payment chain knowledge graph can generate richer semantic representations of nodes, such as density. The importance of a node is measured by the distance between nodes. The shorter the distance from a node to all other nodes, the higher the density, the more central the node is in the network, and the higher its importance.

[0084] The node density is based on the sum of the shortest paths from the given node to all other nodes in the payment chain knowledge graph. Normalization, on the other hand, calculates the average shortest distance from this node to all other nodes. The smaller the average shortest distance of a node, the greater its centrality. If there is no path between node i and node j, then the distance dij from node i to node j is defined as infinity, with its reciprocal being 0.

[0085] In this way, on the one hand, the structured knowledge graph that has been classified and organized can help users intuitively and conveniently explore relationships and perform correlation analysis; on the other hand, it expands the node indicators of the payment chain knowledge graph, including node density and node importance, thereby improving the performance of the model.

[0086] Step S104: The enterprise credit score is calculated based on the node indicators and key payment data using the pre-built final integrated credit assessment model.

[0087] In this embodiment, a final integrated enterprise credit assessment model is constructed based on all relevant key data of the payment chain and node indicators based on the payment chain knowledge graph. The final integrated enterprise credit assessment model is an effective tool to provide correct guidance for decision-making in bank lending. The enterprise credit score calculated based on the final integrated enterprise credit assessment model is relatively accurate, thereby strictly preventing the risks of false transactions and duplicate financing in supply chain finance. At the same time, it can also effectively improve the financing efficiency of upstream and downstream enterprises of core enterprises and solve the problems of difficult financing and high financing costs for small and micro enterprises.

[0088] In one embodiment, constructing the credit assessment model includes:

[0089] An initial integrated credit assessment model is constructed based on multiple initial scenario sub-modules and an initial integration sub-module.

[0090] The supply chain sample data is divided into a training set and a test set;

[0091] The initial integrated credit assessment model is trained using the training set to adjust its parameters, thereby obtaining a modified integrated credit assessment model.

[0092] The modified integrated credit assessment model was tested using the test set, and the test results were obtained.

[0093] The parameters of the modified integrated credit assessment model are adjusted based on the test results to obtain the final integrated credit assessment model.

[0094] In this embodiment, each initial scenario submodule corresponds to a type of scenario data, which can be intra-bank settlement data, intra-bank savings data, inter-bank credit data, order data, invoice data, or payment data. The number of initial scenario submodules is the same as the number of scenario data types. The initial scenario submodules can be constructed using any one of the following four models: random forest, gradient Boosting decision tree (GBDT), XGBoost, or lightGBM. The initial ensemble submodule can use a logistic regression (LG) model. The initial scenario submodules are connected to the initial ensemble submodule; the output data of the initial scenario submodules serves as the input data of the initial ensemble submodules, and the output data of the initial ensemble submodules serves as the output value of the enterprise credit score.

[0095] Before training and testing, supply chain sample data is acquired, which includes sample data from multiple scenarios, such as settlement sample data, intra-bank savings sample data, inter-bank credit information sample data, order sample data, invoice sample data, and payment sample data. The sample data from multiple scenarios are then input into the corresponding initial scenario sub-modules for training, and the parameters of each module are adjusted.

[0096] As an example, the supply chain sample data is divided into a training set and a test set according to a preset ratio, such as 7:3. The training set is divided into K folds to lay the foundation for training each initial scenario sub-module; for example, K=4. Each initial scenario sub-module is trained K times, with 1 / K of the samples retained for validation during each training iteration. After K-fold cross-validation, the predicted values ​​obtained from each fold are used to adjust the parameters of each initial scenario sub-module, resulting in the corrected scenario sub-modules.

[0097] After training, predictions are made on the test set. Each initial scene submodule corresponds to K prediction results, which are then averaged. Finally, the average value of each initial scene submodule after K runs is obtained. The prediction results of each initial scene submodule are concatenated with the real labels of the training set. This concatenated data is then used to train the initial ensemble submodule, adjusting its parameters to obtain the corrected ensemble submodule. Each corrected scene submodule and the corrected ensemble submodule corresponds to a corrected ensemble credit assessment model. Finally, the corrected ensemble credit assessment model is tested on the test set, and the test results are obtained. Based on the difference between the test results and the real labels of the test set, the parameters of the corrected ensemble credit assessment model are adjusted to obtain the final ensemble credit assessment model.

[0098] In one embodiment, the corporate credit assessment method further includes:

[0099] Enterprise tag topics are generated based on the payment chain knowledge graph.

[0100] In this embodiment, the payment chain is the core, acquiring all key payment data related to the transfer of funds between payers and payees. Focusing on enterprise resources, transaction operation information, and external data sources, multiple enterprise tag themes are constructed based on the payment chain knowledge graph. These enterprise tag themes can include basic information, cooperation information, contact information, time information, risk information, value information, financial information, marketing information, tag information, and relationship information. Basic information includes customer overview, customer subsidiary information, shareholder information, etc. Cooperation information includes an overview of customer-held products, customer asset information, and queries for valid corporate customers, etc. Contact information includes customer contact person information and address information, etc. Event information includes customer events, churn warnings, and risk warnings, etc. Risk information includes customer litigation information and social credit information, etc. Value information includes an overview of customer contribution, etc. Financial information includes financial statements, customer asset and balance sheets, etc. Marketing information includes counterparty information and potential customer mining, etc. Tag information includes risk information tags and customer cooperation information tags, etc. Relationship information includes shareholder structure, supply chain relationships, and guarantee relationships, etc.

[0101] In this embodiment, multi-dimensional associated data sources are used, and multiple initial scenario sub-modules are established for different sources, such as savings sub-modules, settlement sub-modules, credit investigation sub-modules, financial sub-modules, payment sub-modules, and tax sub-modules. This effectively captures information from various dimensions and extracts multiple tag themes for corporate clients. Using payment relationships as the core relationship, node indicators of the payment chain knowledge graph are constructed, and the payment chain knowledge graph is visualized.

[0102] The final ensemble credit assessment model provided in this embodiment helps prevent underfitting. By combining all relatively weak features to achieve a complex learning model, it avoids extreme cases, including overfitting. From this perspective, the final ensemble credit assessment model acts as a regularization mechanism, effectively improving the accuracy and robustness of the enterprise credit assessment model presented in this paper. The enterprise credit assessment method provided in this embodiment, while strictly controlling the risks of fraudulent transactions and duplicate financing in supply chain finance and improving the risk management capabilities of supply chain finance, effectively improves the financing efficiency of upstream and downstream enterprises of core enterprises.

[0103] The enterprise credit assessment method provided in this embodiment acquires multi-dimensional supply chain data; preprocesses the multi-dimensional supply chain data to obtain multi-dimensional related data sources; constructs a payment chain knowledge graph based on the multi-dimensional related data sources, and obtains the node indicators of the payment chain knowledge graph; and calculates the enterprise credit score based on the node indicators and key payment data using a pre-constructed final integrated credit assessment model. In this way, by calculating the enterprise credit score using the pre-constructed final integrated credit assessment model based on the node indicators and key payment data of the payment chain knowledge graph, the accuracy of enterprise credit prediction is improved, and the model's generalization ability is enhanced.

[0104] Example 2

[0105] In addition, embodiments of this application provide a corporate credit assessment device.

[0106] Specifically, such as Figure 5 As shown, the enterprise credit assessment device 400 includes:

[0107] Module 401 is used to acquire multidimensional supply chain data;

[0108] Processing module 402 is used to preprocess the multidimensional supply chain data to obtain multidimensional related data sources;

[0109] The construction module 403 is used to construct a payment chain knowledge graph based on the multi-dimensional associated data source and obtain the node indicators of the payment chain knowledge graph.

[0110] The calculation module 404 is used to calculate the enterprise credit score based on the node indicators and payment key data through a pre-built final integrated credit assessment model.

[0111] In one embodiment, the acquisition module 401 is further configured to acquire multiple basic data tables of the supply chain, acquire multidimensional correlation factors of each basic data table, and acquire derived data of each basic data table based on the multidimensional correlation factors;

[0112] The supply chain multidimensional data is composed of multiple basic data tables and multiple derived data tables.

[0113] In one embodiment, the processing module 402 is further configured to preprocess the supply chain multidimensional data using a synthetic minority class oversampling algorithm.

[0114] In one embodiment, the construction module 403 is further configured to acquire multiple knowledge units and obtain the unit relationships between each knowledge unit based on the multi-dimensional associated data source;

[0115] The payment chain knowledge graph is constructed based on the unit relationships between various knowledge units.

[0116] In one embodiment, the construction module 403 is further configured to obtain the sum of distances between each node of the payment chain knowledge graph and other nodes;

[0117] The importance of each node is determined based on its distance from other nodes;

[0118] Obtain the sum of the shortest paths or the average shortest distance between each node and other nodes in the payment chain knowledge graph;

[0119] The node density is determined based on the sum of the shortest paths or the average shortest distance.

[0120] In one embodiment, the construction module 403 is further configured to construct an initial integrated credit assessment model based on multiple initial scenario sub-modules and an initial integration sub-module;

[0121] The supply chain sample data is divided into a training set and a test set;

[0122] The initial integrated credit assessment model is trained using the training set to adjust its parameters, thereby obtaining a modified integrated credit assessment model.

[0123] The modified integrated credit assessment model was tested using the test set, and the test results were obtained.

[0124] The parameters of the modified integrated credit assessment model are adjusted based on the test results to obtain the final integrated credit assessment model.

[0125] In one embodiment, the enterprise credit assessment device 400 further includes:

[0126] The generation module is used to generate enterprise tag themes based on the payment chain knowledge graph.

[0127] The enterprise credit assessment device 400 provided in this embodiment can implement the enterprise credit assessment method provided in Embodiment 1. To avoid repetition, it will not be described again here.

[0128] The enterprise credit assessment device provided in this embodiment acquires multi-dimensional supply chain data; preprocesses the multi-dimensional supply chain data to obtain multi-dimensional related data sources; constructs a payment chain knowledge graph based on the multi-dimensional related data sources, and obtains the node indicators of the payment chain knowledge graph; and calculates an enterprise credit score based on the node indicators and key payment data using a pre-built final integrated credit assessment model. In this way, by calculating an enterprise credit score based on the node indicators and key payment data of the payment chain knowledge graph using a pre-built final integrated credit assessment model, the accuracy of enterprise credit prediction is improved, and the model's generalization ability is enhanced.

[0129] Example 3

[0130] Furthermore, this application provides an electronic device including a memory and a processor. The memory stores a computer program, which executes the enterprise credit assessment method provided in Embodiment 1 when it runs on the processor.

[0131] For details, see Figure 5 The electronic device 500 includes: a transceiver 501, a bus interface, and a processor 502. The processor 502 is used to: acquire multi-dimensional supply chain data.

[0132] The multidimensional supply chain data is preprocessed to obtain a multidimensional related data source;

[0133] A payment chain knowledge graph is constructed based on the multi-dimensional associated data sources, and node indicators of the payment chain knowledge graph are obtained.

[0134] The enterprise credit score is calculated based on the node indicators and key payment data through a pre-built final integrated credit assessment model.

[0135] In one embodiment, the processor 502 is further configured to: acquire multiple basic data tables of the supply chain, acquire multidimensional correlation factors of each basic data table, and acquire derived data of each basic data table based on the multidimensional correlation factors;

[0136] The supply chain multidimensional data is composed of multiple basic data tables and multiple derived data tables.

[0137] In one embodiment, the processor 502 is further configured to: preprocess the supply chain multidimensional data using a synthetic minority class oversampling algorithm.

[0138] In one embodiment, the processor 502 is further configured to: acquire multiple knowledge units, and acquire the unit relationships between the knowledge units based on the multi-dimensional association data source;

[0139] The payment chain knowledge graph is constructed based on the unit relationships between various knowledge units.

[0140] In one embodiment, the processor 502 is further configured to: obtain the sum of distances between each node of the payment chain knowledge graph and other nodes;

[0141] The importance of each node is determined based on its distance from other nodes;

[0142] Obtain the sum of the shortest paths or the average shortest distance between each node and other nodes in the payment chain knowledge graph;

[0143] The node density is determined based on the sum of the shortest paths or the average shortest distance.

[0144] In one embodiment, the processor 502 is further configured to: construct an initial integrated credit assessment model based on multiple initial scenario sub-modules and an initial integration sub-module;

[0145] The supply chain sample data is divided into a training set and a test set;

[0146] The initial integrated credit assessment model is trained using the training set to adjust its parameters, thereby obtaining a modified integrated credit assessment model.

[0147] The modified integrated credit assessment model was tested using the test set, and the test results were obtained.

[0148] The parameters of the modified integrated credit assessment model are adjusted based on the test results to obtain the final integrated credit assessment model.

[0149] In one embodiment, the processor 502 is further configured to: generate enterprise tag topics based on the payment chain knowledge graph.

[0150] In this embodiment of the invention, the electronic device 500 further includes a memory 503. Figure 5 In this context, the bus architecture can include any number of interconnected buses and bridges, specifically linking various circuits together, represented by one or more processors (processor 502) and memory (memory 503). The bus architecture can also link together various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. The bus interface provides an interface. The transceiver 501 can be multiple elements, including transmitters and receivers, providing a unit for communicating with various other devices over a transmission medium. The processor 502 is responsible for managing the bus architecture and general processing, and the memory 503 can store data used by the processor 502 during operation.

[0151] The electronic device 500 provided in this embodiment of the invention can execute the steps of the enterprise credit assessment method provided in the above-described method embodiment 1. To avoid repetition, it will not be described again here.

[0152] The electronic device provided in this embodiment acquires multi-dimensional supply chain data; preprocesses the multi-dimensional supply chain data to obtain multi-dimensional related data sources; constructs a payment chain knowledge graph based on the multi-dimensional related data sources, and obtains the node indicators of the payment chain knowledge graph; and calculates a corporate credit score based on the node indicators and key payment data using a pre-built final integrated credit assessment model. In this way, by calculating a corporate credit score based on the node indicators and key payment data of the payment chain knowledge graph using a pre-built final integrated credit assessment model, the accuracy of corporate credit prediction is improved, and the model's generalization ability is enhanced.

[0153] Example 4

[0154] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the enterprise credit assessment method provided in Embodiment 1.

[0155] In this embodiment, the computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc.

[0156] The computer-readable storage medium provided in this embodiment can implement the enterprise credit assessment method provided in Embodiment 1. To avoid repetition, it will not be described again here.

[0157] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal that includes that element.

[0158] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0159] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A method for assessing corporate credit, characterized in that, The method includes: Acquire multi-dimensional supply chain data; the multi-dimensional supply chain data includes intra-bank settlement data, intra-bank savings data, inter-bank credit data, order data, invoice data, and payment data; The multidimensional supply chain data is preprocessed to obtain a multidimensional related data source; A payment chain knowledge graph is constructed based on the multi-dimensional associated data sources, and node indicators of the payment chain knowledge graph are obtained. The enterprise credit score is calculated based on the node indicators and key payment data through a pre-built final integrated credit assessment model. The node metrics for obtaining the payment chain knowledge graph include: Obtain the sum of the distances between each node and other nodes in the payment chain knowledge graph; The importance of each node is determined based on its distance from other nodes; Obtain the sum of the shortest paths or the average shortest distance between each node and other nodes in the payment chain knowledge graph; The node density is determined based on the sum of the shortest paths or the average shortest distance; Constructing the final integrated credit assessment model includes: An initial integrated credit assessment model is constructed based on multiple initial scenario sub-modules and an initial integration sub-module; the initial scenario sub-modules are connected to the initial integration sub-module, and the output data of the initial scenario sub-modules is used as the input data of the initial integration sub-module; The supply chain sample data is divided into a training set and a test set; The initial integrated credit assessment model is trained using the training set to adjust its parameters, thereby obtaining a modified integrated credit assessment model. The modified integrated credit assessment model was tested using the test set, and the test results were obtained. The parameters of the modified integrated credit assessment model are adjusted based on the test results to obtain the final integrated credit assessment model.

2. The method according to claim 1, characterized in that, The acquisition of multi-dimensional supply chain data includes: Obtain multiple basic data tables of the supply chain, and obtain multidimensional correlation factors for each basic data table; obtain derived data for each basic data table based on the multidimensional correlation factors; The supply chain multidimensional data is composed of multiple basic data tables and multiple derived data tables.

3. The method according to claim 1, characterized in that, The preprocessing of the multidimensional supply chain data includes: The multidimensional data of the supply chain is preprocessed using a synthetic minority class oversampling algorithm.

4. The method according to claim 1, characterized in that, The construction of the payment chain knowledge graph based on the multi-dimensional associated data sources includes: Multiple knowledge units are acquired, and the unit relationships between each knowledge unit are obtained based on the multi-dimensional associated data source. The payment chain knowledge graph is constructed based on the unit relationships between various knowledge units.

5. The method according to claim 1, characterized in that, The method further includes: Enterprise tag topics are generated based on the payment chain knowledge graph.

6. A corporate credit assessment device, characterized in that, The device includes: The acquisition module is used to acquire multi-dimensional supply chain data, which includes intra-bank settlement data, intra-bank savings data, inter-bank credit data, order data, invoice data, and payment data. The processing module is used to preprocess the multidimensional supply chain data to obtain multidimensional related data sources; The construction module is used to construct a payment chain knowledge graph based on the multi-dimensional associated data sources and obtain the node indicators of the payment chain knowledge graph. The calculation module is used to calculate the enterprise credit score based on the node indicators and key payment data through a pre-built final integrated credit assessment model; The node metrics for obtaining the payment chain knowledge graph include: Obtain the sum of the distances between each node and other nodes in the payment chain knowledge graph; The importance of each node is determined based on its distance from other nodes; Obtain the sum of the shortest paths or the average shortest distance between each node and other nodes in the payment chain knowledge graph; The node density is determined based on the sum of the shortest paths or the average shortest distance; Constructing the final integrated credit assessment model includes: An initial integrated credit assessment model is constructed based on multiple initial scenario sub-modules and an initial integration sub-module; the initial scenario sub-modules are connected to the initial integration sub-module, and the output data of the initial scenario sub-modules is used as the input data of the initial integration sub-module; The supply chain sample data is divided into a training set and a test set; The initial integrated credit assessment model is trained using the training set to adjust its parameters, thereby obtaining a modified integrated credit assessment model. The modified integrated credit assessment model was tested using the test set, and the test results were obtained. The parameters of the modified integrated credit assessment model are adjusted based on the test results to obtain the final integrated credit assessment model.

7. An electronic device, characterized in that, It includes a memory and a processor, the memory storing a computer program that executes the enterprise credit assessment method according to any one of claims 1 to 5 when the processor is running.

8. A computer-readable storage medium, characterized in that, It stores a computer program that, when run on a processor, executes the enterprise credit assessment method according to any one of claims 1 to 5.

Images