Enterprise network association risk processing method and device, equipment and medium

Abstract

The invention discloses an enterprise network association risk processing method and device, equipment and a medium. The method comprises the following steps: on the basis of at least one business relationship between enterprises, constructing a weighted financial association network which takes the enterprises as nodes and the business relationships between the enterprises as edges, the weight of the edges being business relationship strength represented by the edges; the business relationship comprises a guarantee relationship, a transaction relationship and an equity relationship; based on the weight of the edge in the weighted financial association network, determining a network distance matrix representing the length of a risk propagation path between any two enterprises; and for a target enterprise, based on the network distance matrix and the basic risk of each risk source of the target enterprise, determining the network association risk of the target enterprise through a risk attenuation model. According to the embodiment of the invention, the enterprise association risk can be determined, and the risk control precision is improved.

Description

Technical Field

[0001] This invention relates to the field of financial technology, and in particular to a method, apparatus, device, and medium for handling enterprise network-related risks. Background Technology

[0002] In recent years, with the rapid development of fintech, artificial intelligence and big data analytics have been increasingly applied in areas such as financial risk management, credit assessment, and robo-advisors. Financial institutions are building complex machine learning models to extract potential patterns from massive amounts of data, aiming to improve the accuracy of risk prediction and the intelligence of business decisions. However, traditional credit scoring models (such as logistic regression and random forests) primarily rely on static financial data of individual companies, neglecting the complex business relationships between companies within the financial network. This limitation makes it difficult for existing models to effectively identify systemic risks, such as cascading defaults triggered by broken guarantee chains or supply chain crises. Furthermore, while existing risk modeling methods based on graph neural networks (GNNs) can capture network correlations, their "black box" nature leads to insufficient model interpretability, making it difficult to meet the transparency requirements of financial regulation. Simultaneously, the high computational complexity of GNNs also limits their application in real-time risk control scenarios. Therefore, there is an urgent need for a method that balances computational efficiency, interpretability, and dynamic adaptability to more accurately characterize risk contagion paths within the financial network and optimize loan portfolios to reduce systemic risk. Summary of the Invention

[0003] This invention provides a method, apparatus, device, and medium for processing enterprise network-related risks, in order to solve the problems of existing credit scoring models neglecting the complex business relationships between enterprises in financial networks and making it difficult to effectively identify systemic risks; as well as the problems of existing risk assessment models having insufficient interpretability, failing to meet the transparency requirements of financial supervision, and having high computational complexity, which limits their application in real-time risk control scenarios.

[0004] According to one aspect of the present invention, a method for handling enterprise network association risks is provided, comprising:

[0005] Based on at least one business relationship between enterprises, a weighted financial relationship network is constructed with enterprises as nodes and business relationships between enterprises as edges, wherein the weight of an edge represents the strength of the business relationship represented by that edge; the business relationships include guarantee relationships, transaction relationships, and equity relationships.

[0006] Based on the weights of the edges in the weighted financial network, a network distance matrix representing the length of the risk propagation path between any two enterprises is determined.

[0007] For the target enterprise, based on the network distance matrix and the basic risks of each risk source of the target enterprise, the network association risk of the target enterprise is determined by the risk attenuation model.

[0008] According to another aspect of the present invention, an apparatus for handling enterprise network association risks is provided, comprising:

[0009] The network construction module is used to construct a weighted financial network based on at least one business relationship between enterprises, with enterprises as nodes and business relationships between enterprises as edges. The weight of an edge represents the strength of the business relationship represented by that edge. The business relationships include guarantee relationships, transaction relationships, and equity relationships.

[0010] The distance matrix construction module is used to determine the network distance matrix representing the length of the risk propagation path between any two enterprises based on the weights of the edges in the weighted financial network.

[0011] The associated risk determination module is used to determine the network associated risk of a target enterprise based on the network distance matrix and the basic risks of each risk source of the target enterprise, and through a risk attenuation model.

[0012] According to another aspect of the present invention, a computer program product is provided, comprising a computer program that, when executed by a processor, implements the method for handling enterprise network association risks according to any embodiment of the present invention.

[0013] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0014] At least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the enterprise network association risk handling method according to any embodiment of the present invention.

[0015] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the enterprise network association risk handling method according to any embodiment of the present invention.

[0016] According to another aspect of the present invention, a computer program product is provided, comprising a computer program / instructions that, when executed by a processor, implement the method for handling enterprise network association risks as described in any embodiment of the present invention.

[0017] This invention constructs a weighted financial relationship network and calculates the network distance matrix to quantify business relationships such as guarantees, transactions, and equity among enterprises into risk propagation paths. Then, based on a risk decay model, it quantifies the associated risks borne by target enterprises due to network connections. This achieves a systematic measurement of indirect and contagious associated risks that are difficult for traditional risk control models to capture, providing a new and quantifiable technical tool for identifying and warning of cascading credit risks caused by broken guarantee chains and supply chain crises. By outputting structured network-related risk quantification results, it directly enhances the existing risk control system's ability to perceive and analyze complex risk networks, fundamentally improving the foresight and comprehensiveness of risk identification.

[0018] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a first flowchart of a method for handling enterprise network association risks provided in an embodiment of the present invention;

[0021] Figure 2 This is a second flowchart of a method for handling enterprise network association risks provided in an embodiment of the present invention;

[0022] Figure 3 This is a schematic diagram of the structure of a device for handling enterprise network association risks provided in an embodiment of the present invention;

[0023] Figure 4 This is a schematic diagram of the structure of an electronic device that implements an embodiment of the present invention. Detailed Implementation

[0024] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0025] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0026] Figure 1 This is a first flowchart of a method for handling enterprise network-related risks provided by an embodiment of the present invention. This embodiment is applicable to scenarios involving early warning of risk contagion in guarantee chains, overcoming the problem that existing credit risk assessment technologies are insufficient in accurately quantifying the risk exposure of individual enterprises within regional enterprise mutual guarantee networks, thus providing early warning of the chain reactions that may be triggered by a single enterprise's default; in supply chain finance risk prevention and control scenarios, addressing the challenge of identifying key nodes and vulnerable links in multi-layered supply chain systems, and assessing the potential impact of supply chain disruption risks on the credit of core enterprises and upstream and downstream enterprises; in enterprise portfolio optimization decision-making scenarios, breaking through the limitations of traditional methods, optimizing credit asset allocation from the perspective of overall network stability, thereby effectively reducing the systemic vulnerability of loan portfolios. This method can be executed by an enterprise network-related risk handling device, which can be implemented in hardware and / or software and can be configured in electronic devices with corresponding data processing capabilities. Figure 1 As shown, the method includes:

[0027] S110. Based on at least one business relationship between enterprises, construct a weighted financial association network with enterprises as nodes and business relationships between enterprises as edges, where the weight of an edge is the strength of the business relationship represented by that edge.

[0028] Based on at least one business relationship between enterprises, a weighted financial relationship network is constructed, with enterprises as nodes and business relationships between enterprises as edges. These relationships include, but are not limited to, guarantee relationships, transaction relationships, and equity relationships. The network supports multiple edges of different types between the same pair of enterprises. Each edge represents a specific type of business relationship between the two enterprises connected by that edge, and each edge is assigned a weight to quantify the strength of the business relationship it corresponds to.

[0029] By constructing this weighted financial relationship network, the originally discrete business relationships between enterprises can be transformed into a structured graph data form, providing a basic data structure for subsequent quantitative analysis of risk propagation based on graph theory. It not only supports network modeling for single business types, but also, through its multi-edge structure, accommodates the coexistence of multiple business relationships between the same pair of enterprises, thus more realistically and comprehensively reflecting the complex relationship patterns between enterprises in actual financial transactions.

[0030] Optionally, the construction of a weighted financial association network based on at least one business relationship between enterprises, with enterprises as nodes and business relationships between enterprises as edges, includes: if there are multiple business relationships between any two enterprises, then multiple independent edges are established between the corresponding node pairs, with each edge corresponding to a business relationship; for each edge, the weight of the edge is calculated according to the type of business relationship corresponding to that edge, using a weight calculation rule adapted to that type.

[0031] Differentiated weighting methods are used to determine the edge weights for various business relationships. Optionally, for guarantee relationships, the guarantee weight is determined based on the guarantee amount and guarantee type as the edge weight; for transaction relationships, the transaction weight is determined based on the transaction amount and transaction frequency as the edge weight; and for equity relationships, the shareholding weight is determined based on the shareholding ratio as the edge weight.

[0032] For guarantee relationships, the guarantee weight is calculated based on the guarantee amount and guarantee type. Guarantee types include joint and several liability guarantees and general guarantee guarantees. The guarantee type coefficient is determined according to the guarantee type. The larger the guarantee amount, the higher the guarantee weight; the guarantee type coefficient of joint and several liability guarantees is greater than that of general guarantee guarantees, as shown in the following formula (1).

[0033] (1)

[0034] The guarantee type coefficient for joint and several liability guarantee is 1.0, and the guarantee type coefficient for general guarantee is 0.7.

[0035] For transaction relationships, the transaction weight is determined based on the transaction amount and transaction frequency to reflect the closeness of the transaction relationship. The transaction amount weight is determined based on the monthly transaction amount, the transaction frequency weight is determined based on the monthly transaction frequency, and the transaction weight is determined based on the transaction amount weight and the transaction frequency weight. The specific formula is shown in formula (2) below.

[0036] (2)

[0037] Where a and b are weighted weights, the sum of a and b is 1, and they are 0.6 and 0.4 respectively.

[0038] All weights are deeply embedded in the banking business logic. For example, when calculating the guarantee weight, the reference amount is set to a benchmark value of 100 million yuan. When calculating the transaction weight, the transaction amount is processed using the natural logarithm to avoid the influence of extreme values. This ensures both the reasonableness of the calculation results and maintains business interpretability.

[0039] Taking a building materials industry cluster as an example, there is a dual relationship between cement plant A and concrete company B: First, cement plant A provides a joint and several liability guarantee of 50 million yuan for concrete company B, with a guarantee weight of 0.5; second, the monthly transaction volume between the two parties is 5 million yuan, with a transaction frequency of 20 times, with a transaction weight of 0.616. The weighted financial network will retain these two independent edges to accurately record the two different types of business relationships.

[0040] By supporting the establishment of multiple independent edges between any two enterprise nodes to represent different types of business relationships (such as guarantees, transactions, equity, etc.), and independently quantifying the relationship strength of each edge using weight calculation rules adapted to that type of business relationship, a more refined and realistic model of financial relationship networks is achieved. This enables weighted financial relationship networks to fully depict the complex business interactions between enterprises, avoiding the simplification and distortion of traditional single-relationship models, thus providing a structurally complete, semantically clear, and weighted data foundation for subsequent quantitative analysis of risk transmission based on distance matrices.

[0041] S120. Based on the weights of the edges in the weighted financial network, determine the network distance matrix that represents the length of the risk propagation path between any two enterprises.

[0042] Optionally, determining the network distance matrix representing the length of the risk propagation path between any two enterprises based on the edge weights in the weighted financial network includes: determining the initial distance between any two directly connected nodes in the weighted financial network based on the edge weights; calculating the shortest weighted path length between any two nodes in the weighted financial network based on the initial distance between all node pairs using a shortest path algorithm; generating a preliminary distance matrix based on the shortest weighted path length between all node pairs, and setting the values ​​of elements in the matrix that exceed a preset effective propagation threshold to infinity to obtain the network distance matrix.

[0043] Based on the weight of each edge in the constructed weighted financial network, a network distance matrix representing the length of the risk propagation path between any two enterprises is calculated and generated.

[0044] For any two directly connected enterprise nodes in the network, the initial distance between the node pair is determined based on the weights of all associated edges between them. The value of the initial distance is negatively correlated with the overall strength of the weights; that is, the closer the business association (the higher the weight), the smaller the corresponding initial distance value, in order to reflect the characteristic that risks are more likely to propagate along strongly associated paths. The specific formula is shown in formula (3) below.

[0045] (3)

[0046] This method ensures that closely related factors correspond to small distance values. Here, weight 1, weight 2, etc., refer to guarantee weights, transaction weights, or equity weights, etc. By converting business relationship weights into initial distances and establishing a correspondence that "the stronger the relationship, the larger the weight, and the shorter the initial distance," the constructed distance matrix can more realistically reflect the impact of the strength of business relationships between enterprises on risk transmission. This method overcomes the shortcomings of traditional graph analysis, which relies solely on hop counts or simple weights, making it difficult to accurately depict the attenuation of risk along the relationship path. Subsequent risk quantification based on the distance matrix better aligns with the actual business principle that "strong relationships are easily transmitted, while weak relationships gradually attenuate," thereby improving the accuracy of risk identification and credit assessment.

[0047] Using the initial distances between all node pairs as basic input, the shortest weighted path length between any two enterprise nodes in a weighted financial network is calculated using a graph theory algorithm for shortest paths from all sources (such as the Floyd-Warshall algorithm). This length represents the cumulative distance of the "shortest effective path" required for risk to propagate from the source enterprise to the target enterprise, reflecting the path of least resistance in risk propagation.

[0048] The shortest weighted path lengths between all node pairs are organized into a preliminary distance matrix. To further optimize efficiency and conform to the attenuation law of risk propagation, the preliminary distance matrix is ​​truncated: elements exceeding a preset effective propagation threshold (default 5.0) are set to infinity, indicating that paths exceeding this threshold are negligible in actual risk transmission. This conforms to the risk attenuation law and improves computational efficiency. After this processing, a compact and business-meaningful network distance matrix is ​​obtained. This matrix serves as the core input for subsequent network association risk calculation and enterprise portfolio optimization, effectively identifying and measuring indirect risk transmission paths that are not directly related but are transmitted through intermediary enterprises, such as indirect risk transmission from enterprise A to enterprise C via enterprise B.

[0049] The calculation principle and steps for using the Floyd-Warshall algorithm to calculate the shortest weighted path length between all enterprise pairs are as follows:

[0050]

[0051] Each element of the initial distance matrix Indicates enterprise To enterprises The risk propagation distance. A smaller distance value means a shorter risk propagation path and a faster transmission speed. For example, the distance from cement plant A to concrete company B is 1.62, while the distance to building material dealer C is 3.40, indicating that the risk propagation of the former is much faster than that of the latter.

[0052] S130. For the target enterprise, based on the network distance matrix and the basic risks of each risk source of the target enterprise, the network association risk of the target enterprise is determined by the risk attenuation model.

[0053] Network-related risk refers to the potential credit loss risk that an enterprise may passively bear through business relationships due to its direct or indirect connection with other enterprises in a weighted financial network (such as a guarantee chain, supply chain, or equity network), which is transmitted from the deterioration of the risk status of related parties. For a target enterprise, the network-related risk borne by the target enterprise is calculated and determined using a risk decay model based on the generated network distance matrix and the basic risk values ​​of each risk source enterprise. Here, the risk source enterprises refer to all other enterprises existing in the same weighted financial network as the target enterprise, and the basic risk value of the risk source enterprise is used to characterize its own credit risk level.

[0054] Optionally, after determining the network association risk of the target enterprise, the method further includes: correcting the basic credit score of the target enterprise based on the network association risk, and determining the corrected credit score of the target enterprise.

[0055] Based on network-related risks, the target company's basic credit score is dynamically adjusted to determine its revised credit score. This revised credit score, while retaining the company's individual credit characteristics, further incorporates the systemic risks it faces in the weighted financial network, achieving an upgrade from individual credit assessment to credit assessment adjusted for network-related risks.

[0056] Optionally, a target combination that meets preset requirements can be selected from the candidate enterprise combination based on the modified credit scores of each enterprise and the network distance matrix.

[0057] Based on the revised credit scores of each enterprise and combined with the network distance matrix, the target portfolio is dynamically selected from the preset candidate enterprise portfolio according to the preset business objectives: when the business objective focuses on maximizing profits, the portfolio with the highest total revised credit scores is selected first; when the business objective focuses on risk control, the portfolio with the largest average network distance between enterprises (i.e. the most dispersed associated risks) is selected first; ultimately, a configurable and interpretable balance decision between profits and risks is achieved, and an enterprise portfolio adapted to different risk preferences is output.

[0058] This invention constructs a weighted financial relationship network and calculates the network distance matrix to quantify business relationships such as guarantees, transactions, and equity among enterprises into risk propagation paths. Then, based on a risk decay model, it quantifies the associated risks borne by target enterprises due to network connections. This achieves a systematic measurement of indirect and contagious associated risks that are difficult for traditional risk control models to capture, providing a new and quantifiable technical tool for identifying and warning of cascading credit risks caused by broken guarantee chains and supply chain crises. By outputting structured network-related risk quantification results, it directly enhances the existing risk control system's ability to perceive and analyze complex risk networks, fundamentally improving the foresight and comprehensiveness of risk identification.

[0059] Figure 2 This is a second flowchart of a method for handling enterprise network association risks provided by an embodiment of the present invention. This embodiment is an optimization and improvement based on the above embodiment. Figure 2 As shown, the method includes:

[0060] S210. Based on at least one business relationship between enterprises, construct a weighted financial association network with enterprises as nodes and business relationships between enterprises as edges, where the weight of an edge is the strength of the business relationship represented by that edge.

[0061] The business relationships mentioned include guarantee relationships, transaction relationships, and equity relationships.

[0062] S220. Based on the weights of the edges in the weighted financial network, determine the network distance matrix that represents the length of the risk propagation path between any two enterprises.

[0063] S230. For the target enterprise, other enterprises in the weighted financial network besides the target enterprise are regarded as risk sources.

[0064] S240. For any risk source, determine the network distance between the target enterprise and the risk source based on the network distance matrix.

[0065] S250. Determine the distance attenuation factor based on the network distance; determine the single association risk of the risk source to the target enterprise based on the basic risk of the risk source, the industry adjustment coefficient of the risk source, and the distance attenuation factor.

[0066] S260. Determine the network association risk of the target enterprise based on the single associated risk of all risk sources of the target enterprise.

[0067] The basic risk of any of the aforementioned risk sources is a quantitative value that characterizes the inherent risk level of the risk source itself, calculated based on the individual static data of the risk source through a preset risk assessment model.

[0068] The calculation process for the single associated risk of any risk source to the target enterprise is shown in the following formula (4).

[0069] (4)

[0070] The basic risk is a quantified value of the inherent risk level of the risk source enterprise, ranging from 0 to 1. The industry adjustment coefficient is a risk amplification / reduction factor for the risk source enterprise based on industry characteristics. The network distance between the target enterprise and the risk source is determined by querying the network distance matrix. The distance attenuation factor is determined based on the network distance between the risk source enterprise and the target enterprise: Distance attenuation factor = , This is the attenuation coefficient. The attenuation coefficient is a key parameter calibrated based on historical data, and its optimal value is determined through regression analysis of historical default data. Industry adjustment coefficient: reflects differences in systemic risk across industries, such as a coefficient of 1.2 for the construction industry (higher risk) and a coefficient of 0.8 for the technology services industry (lower risk).

[0071] The network-related risk of a target company is determined by summing the individual associated risks of all risk sources across all risk sources. This network-related risk comprehensively reflects the target company's risk exposure level within a weighted financial network.

[0072] For example, the single associated risks of the four risk sources of the target enterprise cement plant A are as follows:

[0073] The single association risk between concrete company B and cement plant A is: 0.7 (basic risk) × 1.0 (industry adjustment coefficient) × 0.318 (distance attenuation factor) = 0.223;

[0074] The single association risk of the Provincial Construction Engineering Group with Cement Plant A is: 0.2 × 1.2 × 0.417 = 0.100;

[0075] The single association risk between technology company D and cement plant A is: 0.3 × 0.8 × 0.215 = 0.052;

[0076] The single association risk between building materials distributor C and cement plant A is: 0.6 × 0.9 × 0.091 = 0.049;

[0077] The network association risk of cement plant A is 0.223 + 0.100 + 0.052 + 0.049 = 0.424, indicating that cement plant A faces a moderately high association risk in the weighted financial association network.

[0078] By treating non-target enterprise nodes in the network as risk sources and calculating the risk transmission distance between the risk sources and the target enterprise based on the network distance matrix, and then combining the basic risk value of the risk source, industry adjustment coefficient, and distance-attenuating factor, the single associated risk of each risk source to the target enterprise is quantified, and finally, the network-related risk of the target enterprise is aggregated. This achieves a systematic, traceable, and accurate measurement of the indirect and associated risks borne by enterprises due to complex business networks, significantly improving the ability of traditional credit assessment models to identify systemic risks and cascading default risks, and enhancing their early warning foresight. This makes the risk assessment results more comprehensive and closer to the risk transmission patterns in real business scenarios.

[0079] Optionally, after determining the network association risk of the target enterprise, the method further includes: correcting the basic credit score of the target enterprise based on the network association risk, and determining the corrected credit score of the target enterprise.

[0080] Optionally, the step of correcting the target enterprise's basic credit score based on the network association risk to determine the target enterprise's corrected credit score includes: normalizing the network association risk to obtain a normalized association risk value; calculating a credit score correction coefficient based on the normalized association risk value according to a preset adjustment strength parameter; correcting the basic credit score using the credit score correction coefficient to generate the target enterprise's corrected credit score; wherein the basic credit score is calculated based on the target enterprise's individual static data using a preset credit assessment model; and screening candidate combinations determined from the enterprise subset of the enterprise set based on the corrected credit score. The individual static data includes, but is not limited to, the enterprise's financial statement data, historical credit records, and principal characteristic information.

[0081] The revised credit score of the target company is calculated as shown in the following formula (5):

[0082] (5)

[0083] in: Assess the target company's basic credit score; Risks associated with network connections; To adjust the strength parameter (default 0.15); This indicates that network association risk is normalized to obtain a normalized association risk value, ensuring that the normalized association risk value is within a certain range. Within the range.

[0084] Among them, adjusting the strength parameters To control the impact of network-related risks on credit scores, the optimal value is determined based on historical data backtesting analysis. This indicates a maximum adjustment range of 15%. Network correlation risk is normalized by mapping the network correlation risk R to the [0,1) interval to avoid over-adjustment caused by extreme risk values. ensure The normalized value approaches 1. The adjustment range for different enterprises is proportional to the enterprise's network association risk, thereby achieving dynamic adjustment of the basic credit score of each enterprise.

[0085] For example, if cement plant A has a basic credit score of 750 and a network association risk of 0.424, then The credit score correction factor is Cement Plant A's revised credit score is Cement Plant A's credit score dropped from 750 to 719, but it is still considered a good customer. The adjustment was reasonable, avoiding excessive penalties and providing a quantitative basis for subsequent business portfolio decisions.

[0086] By normalizing network-related risks and dynamically calculating credit score correction coefficients based on preset adjustment strength parameters, the basic credit score of the target enterprise is quantitatively corrected based on these coefficients, resulting in a revised credit score. This integrates the systemic risk borne by enterprises due to network connections into the traditional credit assessment system in an interpretable and controllable manner. Without overturning the original credit assessment logic, it achieves an upgrade from "individual static credit assessment" to "network dynamic credit enhancement," making credit scores more comprehensive, risk-sensitive, and aligned with actual business scenarios. This provides financial institutions with a more scientific and forward-looking decision-making basis in credit approval, risk pricing, and portfolio management.

[0087] Optionally, the step of screening candidate combinations determined from the enterprise subsets of the enterprise set based on the modified credit scores includes: generating multiple enterprise subsets as candidate combinations from the enterprise set based on a preset combination size parameter; calculating in parallel a return index and a risk diversification index for each candidate combination; wherein the return index is determined based on the modified credit scores of the enterprises within the combination, and the risk diversification index is determined based on the network distance between the enterprises within the combination; and screening all candidate combinations based on the return index and the risk diversification index.

[0088] Based on the portfolio size set by the financial institution's actual business, i.e., the preset number of companies in the portfolio, a subset of all companies satisfying this quantity constraint is selected from the weighted financial network of companies, thereby generating multiple candidate portfolio schemes. For each candidate portfolio, a quantitative evaluation is performed from two dimensions: return and risk. In terms of return, the sum of the modified credit scores of all companies in the portfolio is calculated as the portfolio's return indicator, reflecting the overall credit quality and expected return level of the portfolio. In terms of risk, the average network distance between all companies in the portfolio, determined based on the network distance matrix, is used as the portfolio's risk diversification indicator. The risk diversification indicator characterizes the tightness of business connections between companies in the portfolio; a larger average value indicates weaker connections between companies, better risk diversification, and lower systemic risk. Based on the return and risk diversification indicators, all candidate portfolios are screened to determine the target portfolio.

[0089] Based on the calculated return and risk diversification indicators of each candidate portfolio, all candidate portfolios can be screened according to preset decision rules (e.g., pursuing maximum return, minimum risk, or a balance between the two) to determine the target portfolio. This provides financial institutions with a transparent and quantifiable corporate portfolio solution that takes into account both return considerations and systemic risk control.

[0090] For example, in a preset business portfolio with 2 businesses, the revenue metric for the candidate portfolio: Cement Plant A + Concrete Company B is: Cement Plant A's adjusted credit score 717 + Concrete Company B's adjusted credit score 529 = 1246; the risk diversification metric is the average network distance between Cement Plant A and Concrete Company B = 1.62 / 1 = 1.62; the revenue metric and risk diversification metric for each candidate portfolio are as follows:

[0091]

[0092] When the pre-defined decision-making rules prioritize returns or consider the balance between returns and risks, the optimal portfolio of enterprises is determined as: Cement Plant A + Technology Company D. This portfolio achieves the best balance between return and risk objectives, avoiding the risk-concentrated portfolio within the same industry that might be chosen using traditional methods. The enterprise portfolio can also be a loan portfolio used for lending operations.

[0093] By combining modified credit scores with a network distance matrix, a dual-dimensional assessment framework for returns and risk diversification is constructed, upgrading corporate portfolios from simple credit aggregation optimization to systemic risk prevention. It not only quantitatively assesses the overall credit return of the portfolio but also introduces inter-corporate network distance as an indicator of associated risk diversification. This enables financial institutions to intelligently select the optimal portfolio based on preset strategies (such as return priority, risk aversion, or balanced decision-making), effectively mitigating potential systemic risks caused by excessive concentration of business connections between enterprises, and enhancing the scientific rigor, transparency, and adaptability of asset allocation.

[0094] This invention constructs a weighted financial relationship network and calculates the network distance matrix, systematically quantifying multiple business relationships between enterprises, such as guarantees, transactions, and equity, into risk transmission paths. This allows for the accurate identification of multi-source network-related risks faced by target enterprises, achieving a systematic measurement of indirect and contagious related risks that are difficult for traditional risk control models to capture. It provides a novel and quantifiable technical tool for identifying and warning of cascading credit risks caused by broken guarantee chains and supply chain crises. By outputting structured network-related risk quantification results, it directly enhances the existing risk control system's ability to perceive and analyze complex risk networks, fundamentally improving the foresight and comprehensiveness of risk identification.

[0095] Furthermore, it dynamically adjusts the basic credit scores of target enterprises based on network-related risks, ultimately achieving a dual-objective optimization of returns and risk diversification in enterprise portfolio decision-making. This overcomes the limitations of traditional credit assessment models that neglect risk correlation and contagion among enterprises, significantly improving the foresight of risk warnings, the comprehensiveness of credit assessments, and the effectiveness of credit resource allocation. It provides financial institutions with a quantifiable, interpretable, and actionable complete solution for addressing systemic risks in complex network environments.

[0096] Figure 3 This is a schematic diagram of the structure of a device for handling enterprise network association risks provided in an embodiment of the present invention. Figure 3 As shown, the device includes:

[0097] The network construction module 310 is used to construct a weighted financial network based on at least one business relationship between enterprises, with enterprises as nodes and business relationships between enterprises as edges, wherein the weight of an edge is the strength of the business relationship represented by that edge; the business relationship includes guarantee relationship, transaction relationship and equity relationship;

[0098] The distance matrix construction module 320 is used to determine the network distance matrix representing the length of the risk propagation path between any two enterprises based on the weights of the edges in the weighted financial association network.

[0099] The associated risk determination module 330 is used to determine the network associated risk of a target enterprise based on the network distance matrix and the basic risks of each risk source of the target enterprise through a risk attenuation model.

[0100] The enterprise network association risk processing device provided in the embodiments of the present invention can execute the enterprise network association risk processing method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the method execution.

[0101] Optionally, the associated network building modules include:

[0102] An edge establishment unit is used to establish multiple independent edges between corresponding node pairs if there are multiple business relationships between any two enterprises, with each edge corresponding to one business relationship;

[0103] The weight calculation unit is used to calculate the weight of each edge based on the type of business relationship corresponding to that edge, using a weight calculation rule adapted to that type.

[0104] Optional, distance matrix construction module, including:

[0105] The initial distance determination unit is used to determine the initial distance between any two directly connected nodes in the weighted financial network based on the edge weights between each edge.

[0106] The shortest weighted path length determination unit is used to calculate the shortest weighted path length between any two nodes in the weighted financial network based on the initial distance between all node pairs and the shortest path algorithm.

[0107] The distance matrix construction unit is used to generate a preliminary distance matrix based on the shortest weighted path length between all node pairs, and set the values ​​of elements in the matrix that exceed a preset effective propagation threshold to infinity, thereby obtaining the network distance matrix.

[0108] Optionally, the associated risk determination module includes:

[0109] The risk source identification unit is used to identify other companies in the weighted financial network besides the target company as risk sources for the target company.

[0110] The distance attenuation factor determination unit is used to determine the network distance between the target enterprise and the risk source based on the network distance matrix for any risk source; and to determine the distance attenuation factor based on the network distance.

[0111] A single association risk determination unit is used to determine the single association risk of the risk source to the target enterprise based on the basic risk of the risk source, the industry adjustment coefficient of the risk source, and the distance decay factor.

[0112] The network association risk determination unit is used to determine the network association risk of the target enterprise based on the single association risk of all risk sources of the target enterprise; wherein, the basic risk of any of the risk sources is a quantitative value that characterizes the inherent risk level of the risk source itself, calculated based on the individual static data of the risk source through a preset risk assessment model.

[0113] Optionally, a credit correction module may also be included; the credit correction module includes:

[0114] The correction coefficient determination unit is used to normalize the network-related risk to obtain a normalized association risk value; and to calculate the credit score correction coefficient based on the normalized association risk value according to a preset adjustment strength parameter.

[0115] A credit correction unit is used to correct the basic credit score using the credit score correction coefficient to generate a corrected credit score for the target enterprise; wherein the basic credit score is calculated based on the individual static data of the target enterprise using a preset credit assessment model.

[0116] The combined screening unit is used to screen candidate combinations determined from the enterprise subset of the enterprise set based on the modified credit score.

[0117] Optionally, the combined filtering unit includes:

[0118] The candidate combination determination sub-unit is used to generate multiple subsets of enterprises from the enterprise set as candidate combinations based on a preset combination size parameter.

[0119] The return and risk assessment subunit is used to calculate the return index and risk diversification index in parallel for each candidate portfolio; wherein, the return index is determined based on the modified credit scores of the enterprises in the portfolio, and the risk diversification index is determined based on the network distance between the enterprises in the portfolio.

[0120] The combined screening subunit is used to screen all candidate combinations based on the aforementioned return index and risk diversification index.

[0121] The enterprise network association risk processing device described in further detail can also execute the enterprise network association risk processing method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the method.

[0122] According to embodiments of the present invention, the present invention also provides an electronic device, a readable storage medium, and a computer program product.

[0123] Figure 4 A schematic diagram of an electronic device 40 that can be used to implement embodiments of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0124] like Figure 4 As shown, the electronic device 40 includes at least one processor 41 and a memory, such as a read-only memory 42 or a random access memory 43, communicatively connected to the at least one processor 41. The memory stores computer programs executable by the at least one processor. The processor 41 can perform various appropriate actions and processes based on the computer program stored in the read-only memory 42 or loaded from storage unit 48 into the random access memory 43. The random access memory 43 may also store various programs and data required for the operation of the electronic device 40. The processor 41, read-only memory 42, and random access memory 43 are interconnected via a bus 44. An input / output interface 45 is also connected to the bus 44.

[0125] Multiple components in electronic device 40 are connected to input / output interface 45, including: input unit 46, such as keyboard, mouse, etc.; output unit 47, such as various types of monitors, speakers, etc.; storage unit 48, such as disk, optical disk, etc.; and communication unit 49, such as network card, modem, wireless transceiver, etc. Communication unit 49 allows electronic device 40 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0126] Processor 41 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 41 include, but are not limited to, central processing units, graphics processing units, various special-purpose artificial intelligence computing chips, various processors running machine learning model algorithms, digital signal processors, and any suitable processor, controller, microcontroller, etc. Processor 41 performs the various methods and processes described above, such as methods for handling enterprise network association risks.

[0127] In some embodiments, the method for handling enterprise network association risks may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 48. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 40 via read-only memory 42 and / or communication unit 49. When the computer program is loaded into random access memory 43 and executed by processor 41, one or more steps of the enterprise network association risk handling method described above may be performed. Alternatively, in other embodiments, processor 41 may be configured to perform the enterprise network association risk handling method by any other suitable means (e.g., by means of firmware).

[0128] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays, application-specific integrated circuits (ASICs), application-specific standard products (ASICs), systems-on-a-chip (SoCs), payload programmable logic devices, computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0129] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0130] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory, read-only memory, erasable programmable read-only memory, optical fibers, portable compact disk read-only memory, optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0131] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a cathode ray tube, liquid crystal display, or monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0132] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or computing systems that include middleware components (e.g., application servers), or computing systems that include frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0133] A computing system can include clients and servers. Clients and servers are generally geographically separated and typically interact via communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a host product within the cloud computing service system to address the shortcomings of traditional physical hosts and virtual private servers, such as high management difficulty and weak business scalability.

[0134] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0135] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for processing enterprise network association risk, characterized in that, The method comprises: constructing a weighted financial correlation network with enterprises as nodes and business relationships between enterprises as edges based on at least one business relationship between enterprises, wherein the weight of an edge is the strength of the business relationship represented by the edge; the business relationship includes a guarantee relationship, a transaction relationship and a stock relationship; determining a network distance matrix representing the path length of risk transmission between any two enterprises based on the weight of the edge in the weighted financial correlation network; determining the network correlation risk of the target enterprise based on the network distance matrix and the basic risk of each risk source of the target enterprise through a risk attenuation model.

2. The method of claim 1, wherein, The method comprises: if there are multiple business relationships between any two enterprises, multiple independent edges are established between the corresponding node pairs, and each edge corresponds to one business relationship; for each edge, the weight of the edge is calculated according to the type of the business relationship corresponding to the edge using a weight calculation rule suitable for the type.

3. The method of claim 1, wherein, The method comprises: determining the initial distance between the node pair based on the weight of each edge between any two directly connected nodes in the weighted financial correlation network; based on the initial distance between all node pairs, the shortest weighted path length between any two nodes in the weighted financial correlation network is calculated through a shortest path algorithm; based on the shortest weighted path length between all node pairs, a preliminary distance matrix is generated, and the element values in the matrix that exceed a preset effective transmission threshold are set to infinity to obtain the network distance matrix.

4. The method of claim 1, wherein, The method comprises: for the target enterprise, other enterprises in the weighted financial correlation network except the target enterprise are regarded as risk sources; for any risk source, the network distance between the target enterprise and the risk source is determined based on the network distance matrix; a distance attenuation factor is determined according to the network distance; a single correlation risk of the target enterprise to the risk source is determined according to the basic risk of the risk source, the industry adjustment coefficient of the risk source, and the distance attenuation factor; the network correlation risk of the target enterprise is determined according to the single correlation risk of all risk sources of the target enterprise; wherein the basic risk of any risk source is a quantitative value representing the inherent risk level of the risk source itself, which is calculated based on the individual static data of the risk source through a preset risk assessment model.

5. The method of claim 1, wherein, After determining the network correlation risk of the target enterprise, the method further comprises: normalizing the network correlation risk to obtain a normalized correlation risk value; calculating a credit score correction coefficient based on the normalized correlation risk value according to a preset adjustment strength parameter; The base credit score is modified by the credit score modification coefficient to generate a modified credit score of the target enterprise; wherein the base credit score is calculated based on individual static data of the target enterprise by a preset credit evaluation model; The candidate combinations of the enterprise subsets of the enterprise set are screened based on the modified credit score.

6. The method of claim 5, wherein, The screening of the candidate combinations of the enterprise subsets of the enterprise set based on the modified credit score comprises: A plurality of enterprise subsets are generated from the enterprise set as candidate combinations based on a preset combination size parameter; For each candidate combination, its return index and risk dispersion index are calculated in parallel; wherein the return index is determined based on the modified credit scores of the enterprises in the combination, and the risk dispersion index is determined based on the network distances between the enterprises in the combination; All candidate combinations are screened based on the return index and the risk dispersion index.

7. An apparatus for processing of enterprise network associated risks, characterized by The device comprises: An associated network construction module configured to construct a weighted financial associated network with enterprises as nodes and business relationships between enterprises as edges based on at least one business relationship between enterprises, wherein the weight of an edge is the strength of the business relationship represented by the edge; the business relationship includes a guarantee relationship, a transaction relationship and a stock ownership relationship; A distance matrix construction module configured to determine a network distance matrix representing the path length of risk transmission between any two enterprises based on the weight of the edge in the weighted financial associated network; An associated risk determination module configured to determine the network associated risk of a target enterprise based on the network distance matrix and the base risk of each risk source of the target enterprise by a risk attenuation model.

8. An electronic device, comprising: The electronic device comprises: at least one processor; and a memory connected in communication with the at least one processor; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to execute the processing method of the enterprise network associated risk according to any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer instructions for enabling the processor to execute the processing method of the enterprise network associated risk according to any one of claims 1-6 when executed.

10. A computer program product comprising a computer program which, when executed by a processor, implements the processing method of the enterprise network associated risk according to any one of claims 1-6.

Images