Bill transaction risk control method and related equipment

Abstract

The invention relates to a bill transaction risk control method and related equipment, and the method comprises the following steps: carrying out the data mining of the historical transaction data of a bill through a preset data mining technology, and obtaining the transaction behavior characteristics of the bill; performing state space modeling on the bill based on the bill transaction behavior characteristics to obtain a transaction state transition matrix; performing multi-stage risk conduction path simulation on the bill based on the transaction state transition matrix to obtain a risk conduction probability graph; performing risk hotspot identification on the bill based on the risk conduction probability graph to obtain risk hotspot distribution; and carrying out risk coping strategy matching on the bill based on the risk hotspot distribution to obtain a bill risk coping strategy combination, and solving the technical problem that traditional bill risk management mostly depends on static rules and historical experience and is difficult to adapt to a dynamically changing market environment.

Description

Technical Field

[0001] The present invention relates to the technical field of bills, and in particular to a bill transaction risk control method and related equipment. Background Art

[0002] Amidst the increasing complexity and frequency of financial transactions, bills, as a crucial payment and financing tool, promote market liquidity while also introducing significant transaction risks. Traditional bill risk management relies heavily on static rules and historical experience, making it difficult to adapt to the dynamic market environment. This is particularly true in cross-institutional and cross-market transactions, where risk transmission pathways are more subtle and complex. This leads to significant delays in identifying potential risk events and addressing them.

[0003] While existing research attempts to improve bill risk identification capabilities through data analysis and modeling, practical applications still face numerous limitations. For example, most methods focus solely on single-dimensional risk characteristics, lack a comprehensive understanding of the overall dynamic evolution of bill trading behavior, and are unable to effectively capture the nonlinear relationships within the multi-stage risk transmission process. Furthermore, current technical approaches often employ fixed thresholds or simple clustering methods to identify risk hotspots, making it difficult to accurately identify risk clusters and key nodes, thus hindering the accuracy and timeliness of risk warnings.

[0004] At the same time, risk control strategies often lack effective matching mechanisms with specific risk characteristics, resulting in a lack of targeted and flexible strategy combinations. This is particularly true when faced with diverse trading scenarios. The lack of the ability to dynamically generate strategies based on risk transmission paths and hotspot distributions limits the overall effectiveness of bill trading risk prevention and control. Therefore, a comprehensive approach that integrates data mining, state modeling, and risk transmission simulation is urgently needed to achieve systematic identification and intelligent response to bill trading risks. Summary of the Invention

[0005] The main purpose of the present invention is to provide a method for controlling transaction risks of bills, which solves the technical problem that traditional bill risk management relies too much on static rules and historical experience and is difficult to adapt to the dynamically changing market environment.

[0006] To achieve the above object, the present invention provides a method for controlling transaction risks of bills, comprising the following steps: Performing data mining on the historical transaction data of the bill using a preset data mining technology to obtain bill transaction behavior characteristics; Performing state space modeling on the bill based on the bill transaction behavior characteristics to obtain a transaction state transfer matrix; Performing a multi-stage risk transmission path simulation on the bill based on the transaction state transition matrix to obtain a risk transmission probability graph; Based on the risk transmission probability map, risk hotspots of the bill are identified to obtain a risk hotspot distribution; wherein the risk hotspot distribution includes areas where the frequency of risk events is abnormally concentrated, risk transmission intensity peak nodes, and risk concentration centers under different transaction scenarios; Based on the risk hotspot distribution, risk response strategies are matched for the bills to obtain a bill risk response strategy combination.

[0007] Furthermore, the historical transaction data of the bill is mined using a preset data mining technology to obtain bill transaction behavior characteristics, including: Performing transaction behavior structure analysis on historical transaction data of the bill to obtain transaction behavior structure elements, and analyzing structural relationships of the transaction behavior structure elements to obtain transaction behavior structure relationships; Based on the transaction behavior structural relationship, a transaction pattern is constructed for the historical transaction data of the bill to obtain a transaction pattern set, and transaction behavior features in the transaction pattern set are extracted to obtain bill transaction behavior features.

[0008] Furthermore, the state space modeling of the bill based on the bill transaction behavior characteristics to obtain a transaction state transfer matrix includes: Defining state variables for the bill transaction behavior characteristics to obtain bill transaction state variables, and defining a state space range for the bill transaction state variables; Determining a state transition rule for the bill based on the state space range to obtain a bill state transition rule, and setting an initial state of the bill according to the bill state transition rule to obtain an initial transaction state; Simulating a state transition process of the bill based on the initial transaction state to obtain a state transition process, and calculating a state transition probability of the state transition process to obtain a state transition probability; A transaction state transfer matrix is ​​constructed for the bill based on the state transfer probability to obtain a transaction state transfer matrix. Furthermore, the multi-stage risk transmission path simulation of the bill is performed based on the transaction state transition matrix to obtain a risk transmission probability map, including: Dividing the transaction state transfer matrix into risk transmission stages to obtain a multi-stage risk transmission framework, and defining nodes in the multi-stage risk transmission framework to obtain a risk transmission framework with node definitions; Preliminarily setting the risk transmission path of the bill based on the risk transmission framework with node definitions to obtain an initial risk transmission path, and performing a path possibility assessment based on the initial risk transmission path to obtain an assessed risk transmission path; Allocating risk transmission probabilities for the bills based on the assessed risk transmission paths to obtain a risk transmission probability distribution, and adjusting transmission probabilities based on the risk transmission probability distribution to obtain an adjusted risk transmission probability distribution; A risk transmission probability map is constructed for the bill based on the adjusted risk transmission probability distribution to obtain a risk transmission probability map.

[0009] Furthermore, the risk transmission probability distribution of the bill is performed based on the assessed risk transmission path to obtain the risk transmission probability distribution, including: Analyzing the risk transmission correlation factors of the assessed risk transmission path to obtain risk transmission correlation factors, and classifying the risk transmission correlation factors by factor correlation to obtain classified risk transmission correlation factors; Performing an initial assignment of risk transmission probability to the bill based on the classified risk transmission correlation factors to obtain an initial value of the risk transmission probability, and checking the rationality of the initial value of the risk transmission probability to obtain a checked initial value of the risk transmission probability; Dynamically adjusting the risk transmission probability of the bill based on the initial value of the risk transmission probability after the inspection to obtain a dynamically adjusted risk transmission probability, and performing an adjustment rationality assessment on the dynamically adjusted risk transmission probability to obtain a risk transmission probability with an assessed rationality; Based on the risk transmission probability of the rationality of the assessment, a risk transmission probability distribution is constructed for the bill to obtain a risk transmission probability distribution.

[0010] Furthermore, the step of identifying risk hotspots of the bill based on the risk transmission probability map to obtain a risk hotspot distribution includes: Dividing the risk transmission probability map into risk areas to obtain a set of risk areas, and marking risk characteristics of the set of risk areas to obtain risk areas marked with risk characteristics; Performing risk transmission link analysis on the bill based on the risk area with the risk characteristics marked to obtain a risk transmission link set, and performing link risk strength assessment on the risk transmission link set to obtain an assessed link risk strength; Performing preliminary risk hotspot identification on the bill based on the risk intensity of the assessment link to obtain preliminary risk hotspots, and performing hotspot stability analysis based on the preliminary risk hotspots to obtain stable risk hotspots; A risk hotspot distribution is constructed for the bill based on the stable risk hotspot to obtain a risk hotspot distribution.

[0011] Furthermore, the step of performing preliminary risk hotspot identification on the bill based on the risk strength of the assessment link to obtain preliminary risk hotspots includes: Performing risk intensity discretization processing on the risk intensity of the assessment link to obtain a risk intensity discrete value, and defining a discrete value range of the risk intensity discrete value to obtain a defined risk intensity discrete value; Constructing a risk hotspot pre-discrimination standard for the bill based on the defined risk intensity discrete value to obtain a risk hotspot pre-discrimination standard; performing a preliminary risk hotspot identification operation on the bill based on the risk hotspot pre-identification criteria to obtain a preliminary identified risk hotspot, and evaluating the credibility of the preliminary identified risk hotspot to obtain a preliminary identified risk hotspot with an evaluated credibility; Based on the preliminary identified risk hotspots of the assessment credibility, risk correlation factors of the bill are mined to obtain risk hotspot correlation factors, and hotspots of the bill are constructed based on the risk hotspot correlation factors to obtain preliminary risk hotspots.

[0012] The present invention also provides a bill transaction risk control device, comprising: A mining module, configured to perform data mining on the historical transaction data of the bill using a preset data mining technology to obtain bill transaction behavior characteristics; A modeling module, configured to perform state space modeling on the bill based on the bill transaction behavior characteristics to obtain a transaction state transfer matrix; a simulation module, configured to simulate a multi-stage risk transmission path of the bill based on the transaction state transition matrix to obtain a risk transmission probability graph; an identification module for identifying risk hotspots for the bill based on the risk transmission probability map to obtain a risk hotspot distribution; wherein the risk hotspot distribution includes areas where the frequency of risk events is abnormally concentrated, risk transmission intensity peak nodes, and risk concentration centers under different transaction scenarios; A matching module is used to match the risk response strategy of the bill based on the risk hotspot distribution to obtain a bill risk response strategy combination.

[0013] The present invention also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the above methods when executing the computer program.

[0014] The present invention also provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of any of the above methods are implemented.

[0015] The present invention provides a transaction risk control method for bills, comprising the following steps: performing data mining on historical transaction data of the bills by using a preset data mining technology to obtain bill transaction behavior characteristics; performing state space modeling on the bills based on the bill transaction behavior characteristics to obtain a transaction state transfer matrix; performing multi-stage risk transmission path simulation on the bills based on the transaction state transfer matrix to obtain a risk transmission probability map; identifying risk hotspots on the bills based on the risk transmission probability map to obtain a risk hotspot distribution; matching risk response strategies for the bills based on the risk hotspot distribution to obtain a bill risk response strategy combination, which solves the technical problem that traditional bill risk management relies too much on static rules and historical experience and is difficult to adapt to a dynamically changing market environment, realizes the construction of a transaction state transfer matrix based on a state space model, can quantitatively model the state changes of bills in different transaction stages, capture their dynamic evolution laws, and thus improve the technical effect of the interpretability and predictability of risk transmission paths in the bill transaction process. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a schematic diagram of the steps of a method for controlling transaction risks of bills in one embodiment of the present invention; Figure 2 This is a structural block diagram of a bill transaction risk control device according to an embodiment of the present invention; Figure 3 It is a schematic block diagram of the structure of a computer device according to an embodiment of the present invention.

[0017] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0019] like Figure 1 As shown, Figure 1 A method for controlling transaction risks of bills in one embodiment of the present invention includes the following steps: Step S1, performing data mining on the historical transaction data of the bill using a preset data mining technology to obtain bill transaction behavior characteristics.

[0020] Specifically, historical bill transaction data is mined using pre-defined data mining techniques to identify bill transaction behavior characteristics. This process is a fundamental component of the overall risk control approach, aiming to extract quantitative indicators that reflect the patterns and potential risk characteristics of bill transactions from a large volume of historical transaction records. In practical implementation, a unified data processing framework must first be established to cleanse and standardize bill transaction data (such as transaction time, transaction amount, transaction subject, circulation path, and redemption status) from various sources and formats to ensure data quality for subsequent analysis. Subsequently, pre-defined data mining techniques, such as cluster analysis, association rule mining, sequential pattern discovery, or anomaly detection algorithms, are applied to this data to identify typical and abnormal behavior patterns of bills in different transaction scenarios, thereby generating statistically significant and business-interpretable bill transaction behavior characteristics. For example, in the interbank bill discount market, the system can identify behavioral characteristics of bills that frequently circulate between specific institutions by mining association rules based on historical transaction frequency, transaction intervals, and the distribution of transaction parties' credit ratings. This provides a basis for subsequent state modeling and risk transmission analysis. This step not only improves the data quality of risk identification, but also provides support for subsequent risk modeling based on behavioral characteristics.

[0021] Step S2: performing state space modeling on the bill based on the bill transaction behavior characteristics to obtain a transaction state transfer matrix.

[0022] Specifically, state-space modeling of the bills based on their transaction behavior characteristics to generate a transaction state transition matrix is ​​a key step in transforming the transaction behavior characteristics extracted in the previous step into a structured representation of states useful for dynamic risk analysis. This process abstracts and defines various possible "states" that a bill may occupy during a transaction, such as "normal circulation," "high-frequency rediscount," "approaching redemption," or "credit rating downgrade," with clear business implications. Combined with the behavioral characteristics exhibited by the bills at different time points, a discrete state-space model is constructed to characterize the state evolution path of the bills over their transaction lifecycle. Based on this, statistical learning methods or Markov modeling techniques are further utilized to calculate the transition frequencies and probabilities between different states, ultimately forming a transaction state transition matrix that quantifies the state evolution patterns of the bills under various transaction behaviors. For example, in the interbank bill discount market, if a certain type of bill frequently transitions from "normal circulation" to "abnormal high-frequency trading," the corresponding state transition probability will be significantly amplified, reflecting potential risk accumulation trends and providing fundamental support for subsequent multi-stage risk transmission path simulations.

[0023] Step S3: simulate the multi-stage risk transmission path of the bill based on the transaction state transition matrix to obtain a risk transmission probability map.

[0024] Specifically, based on the transaction state transition matrix, a multi-stage risk transmission path simulation is performed on the bill to generate a risk transmission probability map. This is achieved by extending the state transition relationship constructed in the previous step to multiple time stages or transaction nodes, thereby simulating the possible propagation paths and evolution trends of risk events for the bill under different trading environments. In specific implementation, an initial risk state is set as the transmission starting point based on the transition probabilities between each state described by the transaction state transition matrix. For example, the bill is in the "abnormal high-frequency trading" state. Subsequently, the risk diffusion from this initial state to other states is gradually deduced chronologically or at the transaction level. The cumulative risk intensity and transmission probability along each potential transmission path are calculated, ultimately forming a risk transmission network structure covering multiple stages and multiple nodes, namely the risk transmission probability map. This map not only reflects the possibility of risk evolution between different transaction states but also reveals the amplification effect of certain key intermediate links in the transmission process. For example, in the interbank bill discount market, if a bill enters a "high-risk state" at a certain stage due to the deterioration of the credit of the trading entity, the system can track its subsequent circulation path through multi-stage simulation and identify which status nodes are most likely to become risk gathering points, thereby providing a visual and quantitative analysis basis for subsequent risk hotspot identification.

[0025] Step S4: Identify risk hotspots of the bill based on the risk transmission probability map to obtain a risk hotspot distribution; wherein the risk hotspot distribution includes areas where the frequency of risk events is abnormally concentrated, risk transmission intensity peak nodes, and risk concentration centers under different transaction scenarios.

[0026] Specifically, based on the risk transmission probability graph, the risk hotspots of the bills are identified to obtain a risk hotspot distribution. This is achieved by mining nodes or regions with significant risk characteristics in the risk transmission network generated by the aforementioned simulation to accurately locate potential high-risk links in the bill transaction process. In specific implementation, graph analysis technology is first used to calculate the topological properties of the node degree, centrality, path weight, etc. in the risk transmission probability graph, and combined with statistical detection methods, clusters where the frequency of risk events is significantly higher than the average level are identified. At the same time, intensity peak nodes that play a key role in the transmission process are identified. These nodes often represent "hubs" or "amplifiers" in the risk transmission process. In addition, in different trading scenarios, such as periods of tight market liquidity or outbreaks of credit defaults, the system can further identify risk concentration centers in specific scenarios, and ultimately integrate these identification results into a risk hotspot distribution containing multi-dimensional risk characteristics. For example, in the interbank bill discount market, if a certain type of bill frequently appears at key nodes in the risk transmission path, and the trading entities in which it is located exhibit high risk diffusion capabilities in multiple transmission paths, then this type of bill and its associated nodes will be marked as risk hotspots, thereby providing a core basis for subsequent strategy matching. The risk hotspot distribution includes areas with abnormal frequency of risk events, peak nodes of risk transmission intensity, and risk concentration centers under different trading scenarios.

[0027] Step S5: matching the risk response strategies of the bills based on the risk hotspot distribution to obtain a bill risk response strategy combination.

[0028] Specifically, risk response strategies are matched against the bills based on the distribution of risk hotspots to generate a bill risk response strategy combination. This involves intelligently matching the risk event frequency clusters, risk transmission intensity peaks, and risk concentration centers identified in the previous step with a pre-set risk response strategy library, thereby generating differentiated risk control solutions tailored to the specific characteristics of bill transactions. This process constructs a strategy matching model, combines the type and intensity of the risk hotspot, and the trading environment in which it occurs, and utilizes a rules engine or machine learning algorithm to automatically select the most appropriate response measures. These measures are then combined into a multi-level, multi-dimensional strategy combination to effectively mitigate risk. For example, in the interbank bill discount market, if a bill is identified as being in a hotspot with "high frequency of circulation and medium redemption risk," the system can automatically match strategies such as increasing the pledge ratio, restricting the scope of trading partners, and strengthening credit review, forming a targeted and dynamically adaptable bill risk response strategy combination, thereby enhancing overall risk prevention and control capabilities and ensuring transaction security.

[0029] In a specific embodiment, the data mining of the historical transaction data of the bill using a preset data mining technology to obtain bill transaction behavior characteristics includes: Performing transaction behavior structure analysis on historical transaction data of the bill to obtain transaction behavior structure elements, and analyzing structural relationships of the transaction behavior structure elements to obtain transaction behavior structure relationships; Based on the transaction behavior structural relationship, a transaction pattern is constructed for the historical transaction data of the bill to obtain a transaction pattern set, and transaction behavior features in the transaction pattern set are extracted to obtain bill transaction behavior features.

[0030] Specifically, mining the historical transaction data of the bills using pre-defined data mining techniques to obtain bill transaction behavior characteristics is one of the most fundamental and critical steps in the entire bill transaction risk control method. The core of this step lies in extracting behavioral characteristics that reflect the patterns and potential risk characteristics of bill transactions through in-depth analysis of historical transaction data, providing high-quality data support for subsequent state modeling, risk transmission simulation, and hotspot identification. In its implementation, this step includes two key sub-processes: first, analyzing the transaction behavior structure of the historical transaction data to obtain transaction behavior structure elements, and further analyzing the structural relationships between these transaction behavior structure elements to obtain transaction behavior structure relationships; second, constructing transaction patterns from the historical transaction data based on these structural relationships to form a set of transaction patterns, and extracting representative transaction behavior characteristics from them, ultimately generating bill transaction behavior characteristics. In the first step, the system starts with raw bill transaction data. This data typically originates from the interbank bill discount market, inter-enterprise bill circulation platforms, or the bill management systems within financial institutions. It covers multiple dimensions such as transaction time, transaction amount, identities of the two parties to the transaction, bill type, face value, discount rate, maturity date, redemption status, and credit rating. For example, a commercial bank processed over 500,000 bill transactions over the past year, each containing dozens of fields. To more effectively utilize this data, the system needs to perform "transaction behavior structure analysis." This involves abstracting key variables from the raw data into quantifiable and modelable "transaction behavior structure elements" through information extraction, field classification, and logical association. For example, during this process, the system might extract indicators such as "transaction frequency per unit time," "transaction object neutrality index," "average discount rate fluctuation," and "bill life cycle stage distribution" as transaction behavior structure elements, ultimately forming a complete "transaction behavior structure element." Next, the system conducts an in-depth analysis of the structural relationships within these transaction behavior structure elements to reveal the mutual influence mechanisms and evolutionary patterns between different elements, thereby obtaining "transaction behavior structure relationships." For example, the system can use methods such as correlation analysis, principal component analysis (PCA), or network graph modeling to identify strong correlations between transaction behavior structure elements. Hypothetical analysis results show a significant positive correlation between the "frequency of transactions per unit time" and the "transaction target median index" (correlation coefficient reached 0.78). This suggests that certain bills may be traded repeatedly by a small number of institutions within a short period of time, forming a high-frequency, low-dispersion trading pattern. This type of behavior is often associated with arbitrage or abnormal capital flows, and is a potential risk signal. Through this analysis, the system can establish dynamic connections between the various structural elements of trading behavior, laying the foundation for the next step of constructing a set of trading patterns.In the second step, based on the aforementioned transaction behavior structural relationships, the system constructs transaction patterns from historical bill transaction data, generating a set of transaction patterns. The core of this process is to organize previously discrete and fragmented transaction behaviors into a number of representative and business-relevant transaction patterns. For example, the system can group large numbers of bill transaction records using clustering algorithms (such as K-means and DBSCAN) or sequential pattern mining techniques (such as PrefixSpan) to identify several common transaction behavior patterns, such as "high-frequency rediscount patterns," "long-term holding patterns," "near-payment transfer patterns," and "cross-institutional circular transaction patterns." These transaction patterns not only reflect the bill's market circulation path and the behavioral preferences of trading entities, but also provide a structured basis for the subsequent extraction of transaction behavior features. The system then extracts key indicators from these transaction patterns that characterize bill transaction behavior, forming "bill transaction behavior features." These features typically include, but are not limited to, the average bill turnover period, the number of counterparties, the standard deviation of transaction frequency, the discount rate deviation, the number of transactions within 30 days prior to redemption, and the presence of high-risk counterparties. For example, in a bank's 2024 bill trading data, the system identified a certain type of bill with an average of 6.2 transactions within the 30 days prior to maturity, significantly exceeding the average for similar bills (1.5 transactions). Furthermore, several of these counterparties had credit ratings below A. This behavioral characteristic suggests that these bills may be frequently transferred before redemption, attempting to shift risk, a typical high-risk behavior. In summary, by analyzing the transaction behavior structure of historical bill transaction data, extracting transaction behavior structural elements and analyzing their structural relationships, constructing a set of transaction patterns based on these structural relationships, and extracting transaction behavior characteristics from these patterns, the system can comprehensively and systematically characterize the inherent patterns and potential risks of bill trading behavior. This process not only transforms raw transaction data into structured transaction behavior characteristics but also provides solid data support and technical assurance for subsequent transaction status modeling, risk transmission path simulation, and risk response strategy matching. In practical application in the interbank bill discount market, this approach has successfully helped multiple financial institutions proactively identify hundreds of potential illegal transactions, enhancing the risk prevention and control capabilities of the overall bill trading system.

[0031] In a specific embodiment, the state space modeling of the bill based on the bill transaction behavior characteristics to obtain a transaction state transition matrix includes: Defining state variables for the bill transaction behavior characteristics to obtain bill transaction state variables, and defining a state space range for the bill transaction state variables to obtain a defined state space range; Determining a state transition rule for the bill based on the defined state space range to obtain a bill state transition rule, and setting an initial state of the bill according to the bill state transition rule to obtain an initial transaction state; Simulating a state transition process of the bill based on the initial transaction state to obtain a state transition process, and calculating a state transition probability of the state transition process to obtain a state transition probability; A transaction state transfer matrix is ​​constructed for the bill based on the state transfer probability to obtain a transaction state transfer matrix. Specifically, the system first receives a set of bill transaction behavior characteristics output by the preceding module. This set includes, but is not limited to, indicators such as transaction frequency per unit time, counterparty concentration index, discount rate fluctuation, number of transfers within a specific period before redemption, and counterparty credit rating distribution. These indicators reflect the behavioral characteristics of bills at different transaction stages and their potential risk exposure. The system then defines state variables based on this set of bill transaction behavior characteristics. This process maps the raw behavioral characteristics to state labels with clear business meanings, forming a set of bill transaction state variables. For example, states such as "normal circulation," "high frequency rediscount," "approaching redemption," and "credit rating downgrade" are defined as key states that a bill may experience during a transaction. Each state corresponds to a set of quantifiable thresholds or ranges used to determine whether a bill is in that state. For example, the condition for the "high frequency rediscount" state could be set as: the bill has been traded more than five times in the last 30 days; the "approaching redemption" state is defined as a bill with less than 30 days remaining until maturity and has not been redeemed. On this basis, the system further defines the state space scope of the bill transaction state variable set, defining the value boundaries of each state variable to ensure enforceability and consistency. For example, the transaction frequency range for the "normal circulation" state is set to 1 to 3 times per month, and the "credit rating downgrade" state is limited to situations where the counterparty's credit rating drops from A or above to BBB or below. Next, the system establishes a set of bill state transition rules based on the actual bill transaction process and historical data statistics. This rule set describes the trigger conditions and evolutionary paths for bill transitions between different states. For example, if the transaction frequency of a bill in the "normal circulation" state exceeds a set threshold for two consecutive cycles, a transition to the "high-frequency rediscount" state is triggered. If one of its counterparties has a credit rating below B, the bill may enter the "high-risk transaction" state. The system also sets the bill's initial transaction state based on actual transaction scenarios. For example, newly issued bills are typically initialized to "normal circulation," while bills with frequent changes in counterparties are set to "high-frequency rediscount." After obtaining the initial transaction state, the system simulates the state transition process to generate a state transition process set. This process simulates the state transition paths that a bill may undergo over multiple trading cycles, generating multiple typical state transition trajectories. For example, in the interbank bill discount market, a bill may begin in "normal circulation" and progress through multiple states, such as "high-frequency rediscount," "credit rating downgrade," and "high-risk transaction," ultimately entering the "approaching redemption period" state. The system repeatedly generates a large number of state transition paths using methods such as Monte Carlo simulation to enhance the model's generalization and adaptability. The system then statistically analyzes the state transition frequencies along each path in the state transition process set, calculating the transition probabilities between each state and forming a state transition probability set.For example, the average transition probability from the "normal circulation" state to the "high-frequency rediscount" state is 0.32, indicating a roughly one-third probability of such a state jump. The transition probability from "high-frequency rediscount" to "high-risk transaction" is 0.45, indicating a strong risk escalation trend. Ultimately, the system constructs a complete transaction state transition matrix based on the state transition probability set. This matrix uses state variables as rows and columns, with matrix elements representing the probability of transitioning from one state to another. This enables structured modeling of the evolution of bill transaction states. For example, the transaction state transition matrix for a particular type of bill might include four primary states: "normal circulation," "high-frequency rediscount," "approaching redemption," and "high-risk transaction." The matrix elements reflect the evolutionary relationships between these states. Using this matrix, the system can identify which state transitions occur most frequently and which paths are more likely to lead to risk accumulation, providing key input for subsequent multi-stage risk transmission path simulations. For example, in a commercial bank's bill trading system, the system, through state-space modeling, identified a typical evolutionary path for a type of bill: "normal circulation → high-frequency rediscount → credit rating downgrade → high-risk transaction → redemption default." The transition probability from "high-frequency rediscount" to "credit rating downgrade" reached 0.68, indicating a high likelihood of risk evolution along this path. This finding provides strong support for subsequent risk hotspot identification and strategy matching. In summary, by defining state variables and delimiting the state space for the characteristic set of bill trading behavior, formulating state transition rules and setting initial states, simulating the state transition process and calculating transition probabilities, and ultimately constructing a transaction state transition matrix, the entire process achieves structured modeling and dynamic evolution characterization of bill trading states. This method has been deployed and applied in the risk management systems of multiple financial institutions, helping them effectively identify and precisely control bill trading risks, thereby improving the risk control capabilities and operational efficiency of the overall bill market.

[0032] In a specific embodiment, the multi-stage risk transmission path simulation of the bill based on the transaction state transition matrix to obtain a risk transmission probability map includes: Dividing the transaction state transfer matrix into risk transmission stages to obtain a multi-stage risk transmission framework, and defining nodes in the multi-stage risk transmission framework to obtain a risk transmission framework with node definitions; Preliminarily setting the risk transmission path of the bill based on the risk transmission framework with node definitions to obtain an initial risk transmission path, and performing a path possibility assessment based on the initial risk transmission path to obtain an assessed risk transmission path; Allocating risk transmission probabilities for the bills based on the assessed risk transmission paths to obtain a risk transmission probability distribution, and adjusting transmission probabilities based on the risk transmission probability distribution to obtain an adjusted risk transmission probability distribution; A risk transmission probability map is constructed for the bill based on the adjusted risk transmission probability distribution to obtain a risk transmission probability map.

[0033] Specifically, simulating the multi-stage risk transmission paths of the bills based on the transaction state transition matrix to generate a risk transmission probability map is a key step in the overall bill transaction risk control method for dynamic risk propagation analysis and visualization. By introducing the temporal dimension of "multi-stage," this step expands the originally static transaction state transition relationship into multiple evolutionary stages. Combining mechanisms such as node definition, path setting, and probability assignment, a risk transmission probability map is constructed that reflects the gradual evolution, diffusion, and aggregation of risk during the bill transaction process. The specific implementation involves four core steps: first, dividing the transaction state transition matrix into risk transmission stages to form a multi-stage risk transmission framework, based on which nodes are defined to obtain a risk transmission framework with node definitions; second, preliminarily defining risk transmission paths for the bills based on this framework to generate initial risk transmission paths, and then conducting a probability assessment on these paths to select the assessed risk transmission paths; third, assigning risk transmission probabilities to these paths based on the assessment results to form a risk transmission probability distribution, which is then adjusted based on actual data feedback or business rules to obtain an adjusted risk transmission probability distribution; and fourth, constructing a complete risk transmission probability map based on these adjusted probability distributions. First, when stratifying the transaction state transition matrix into risk transmission stages, the system divides the entire bill circulation process into several consecutive and mutually exclusive stages based on the temporal characteristics of the bill transaction cycle and the typical time points at which risk events may occur. For example, in the interbank bill discount market, the bill lifecycle can be divided into four main stages: "initial post-issuance," "mid-term circulation," "approaching the redemption period," and "post-redemption settlement." Each stage corresponds to different trading behavior characteristics and risk exposure levels. The system then defines key risk transmission nodes within each stage, such as "first discount institution," "high-frequency rediscount participant," "credit rating downgrade adjustment point," and "redemption default trigger point," thereby forming a risk transmission framework with node definitions. These nodes not only represent the bill's status at different stages but also reflect its position and influence in the trading network. Next, based on this node-defined risk transmission framework, the system preliminarily defines the risk transmission path for the bill. Specifically, the system uses the transaction state transition matrix constructed in the previous step as the basic rule for path evolution. Starting from the initial transaction state, the system simulates the possible risk transmission paths that the bill may take at different stages based on the transition probabilities between each state. For example, a bill might start in "normal circulation," enter "high-frequency rediscount" in the second stage, enter "high-risk transaction" in the third stage due to a counterparty credit rating decline, and ultimately default in the fourth stage. This creates a complete path: "normal circulation → high-frequency rediscount → high-risk transaction → default." The system generates hundreds or even thousands of such paths through extensive simulations, forming the initial risk transmission path.To ensure the validity and representativeness of the path, the system also assesses its likelihood. Criteria include the path's overall transition probability, its conformity to historical trading patterns, and the presence of unusual transitions. For example, if a bill in a path transitions directly from "normal circulation" to "payment default" between two adjacent stages, and the cumulative transition probability for that path is less than 0.01, the path is deemed low-likely and eliminated. The remaining path becomes the assessed risk transmission path. The system then assigns risk transmission probabilities based on the assessed risk transmission paths. The core of this process is to quantitatively map each node on each path to its corresponding risk transmission strength. For example, for a path like "normal circulation → high-frequency rediscount → high-risk transaction → payment default," the system assigns a risk transmission probability value to each node (e.g., "high-frequency rediscount"), representing the node's contribution to risk propagation along the entire path. This probability assignment is calculated based on factors such as the node's positional weight within the path, the frequency of transitions between nodes, and the node's correlation with other risk events in historical data. For example, the "High-Frequency Rediscount" node, which amplifies risk throughout the entire path, can have its transmission probability set to 0.62. Meanwhile, the "High-Risk Trading" node, assuming it is in a critical state before a risk outbreak, can have its transmission probability set to 0.85. Furthermore, the system dynamically adjusts these probability values ​​based on external data feedback (such as recent market fluctuations and regulatory policy changes) to adapt to the evolving trading environment, thereby forming an adjusted risk transmission probability distribution. Finally, based on this adjusted risk transmission probability distribution, the system constructs a risk transmission probability graph. This graph uses nodes as vertices and paths as edges. The value on each edge represents the probability of risk transmission along that path, and the connections between nodes reflect the direction and intensity of risk transmission across different stages and states. For example, in the interbank bill discount market, a risk transmission probability graph for a particular bill might display the following structure: Initially in the "normal circulation" state, the probability of transmission to the "high-frequency rediscount" state is 0.62; the probability of transmission from "high-frequency rediscount" to "high-risk transaction" is 0.73; and the probability of transmission from "high-risk transaction" to "payment default" is 0.89. This graph clearly illustrates the gradual escalation of risk during the bill transaction process and reveals which nodes are most likely to become "hotspots" for risk transmission. In summary, by expanding the transaction state transition matrix into a multi-stage risk transmission framework and incorporating technical techniques such as node definition, path setting, probability allocation, and adjustment, the system can comprehensively characterize the dynamic evolution of risk during bill transactions and construct a visual risk transmission probability graph. This approach has been deployed in the actual risk control systems of multiple financial institutions, helping them identify hundreds of potential high-risk bill transaction paths, significantly improving the accuracy and timeliness of bill transaction risk warnings.

[0034] In a specific embodiment, the risk transmission probability allocation for the bill based on the assessed risk transmission path to obtain the risk transmission probability allocation includes: Analyzing the risk transmission correlation factors of the assessed risk transmission path to obtain risk transmission correlation factors, and classifying the risk transmission correlation factors by factor correlation to obtain classified risk transmission correlation factors; Performing an initial assignment of risk transmission probability to the bill based on the classified risk transmission correlation factors to obtain an initial value of the risk transmission probability, and checking the rationality of the initial value of the risk transmission probability to obtain a checked initial value of the risk transmission probability; Dynamically adjusting the risk transmission probability of the bill based on the initial value of the risk transmission probability after the inspection to obtain a dynamically adjusted risk transmission probability, and performing an adjustment rationality assessment on the dynamically adjusted risk transmission probability to obtain a risk transmission probability with an assessed rationality; Based on the risk transmission probability of the rationality of the assessment, a risk transmission probability distribution is constructed for the bill to obtain a risk transmission probability distribution. Specifically, assigning risk transmission probabilities to the bills based on the assessed risk transmission pathways is a crucial step in achieving quantitative risk propagation modeling during the multi-stage risk transmission pathway simulation process. This step, through in-depth analysis of the risk transmission-related factors contained in the assessed risk transmission pathways and combining mechanisms such as classification, initial assignment, dynamic adjustment, and rationality assessment, constructs a business-relevant and data-supported risk transmission probability allocation system for bills. Its implementation process includes four key stages: first, analyzing the risk transmission-related factors in the assessed risk transmission pathways to generate risk transmission-related factors, and then classifying them based on their relevance to obtain classified risk transmission-related factors; second, assigning initial risk transmission probabilities to the bills based on these classified factors to generate initial risk transmission probability values, which are then checked for rationality to form the checked initial risk transmission probability values; third, dynamically adjusting the risk transmission probability based on these checked initial values ​​to obtain the dynamically adjusted risk transmission probability, and further evaluating the rationality of the adjustments to form the assessed rationality risk transmission probability; and fourth, ultimately constructing a complete risk transmission probability allocation based on this foundation. In the first step, the system conducts an in-depth analysis of the assessed risk transmission pathways, identifying key factors influencing the intensity of risk transmission during bill transactions. These factors typically include, but are not limited to, "transaction frequency," "transaction counterparty credit rating," "discount rate deviation," "bill lifecycle stage," "market liquidity," and "historical default record." For example, in the interbank bill discount market, the system discovered that a certain type of bill frequently appeared in high-risk transaction pathways in the second phase. These common characteristics were the presence of three counterparties with credit ratings below B, and an average discount rate that was 1.2 percentage points higher than the market average for the same period. These characteristics were grouped into a set of variables with a high correlation with risk transmission, forming risk transmission correlation factors. The system then categorized these factors by correlation, for example, grouping "transaction frequency" and "moderate transaction counterparty" into one category (denoted as "high-frequency trading"), "credit rating" and "historical default rate" into another (denoted as "credit risk"), and "discount rate fluctuation" and "face value deviation" into a third category (denoted as "price anomaly"), thus forming the classified risk transmission correlation factors. Next, the system assigns an initial risk transmission probability to the bill based on the risk transmission-related factors classified above. This process relies on a combination of historical data analysis and expert experience.For example, for bill paths classified as "high-frequency trading," the system, based on historical data from the past three years, found that such paths have a 68% probability of triggering a subsequent high-risk state jump. Therefore, it sets an initial transmission probability of 0.65. For "credit risk" paths, the system sets an initial transmission probability of 0.72 based on the proportion of default cases disclosed in regulatory reports. And for "price anomaly" paths, the system sets an initial probability of 0.58 based on their deviation from the market benchmark. The initial transmission probabilities for all paths are aggregated to form the initial risk transmission probability value. The system then conducts a rationality check on these initial values, primarily by comparing them with similar bill paths, comparing them with industry averages, and verifying them with model predictions. For example, if the initial transmission probability for a path is set at 0.92, but its actual behavioral characteristics do not differ significantly from similar paths, the system adjusts it to 0.78 to avoid overestimating the risk transmission strength, thus forming the initial value of the verified risk transmission probability. The system then dynamically adjusts the initial probability based on this verified initial value by incorporating external real-time data sources and feedback mechanisms. For example, in the fourth quarter of 2024, if liquidity constraints at financial institutions in a certain region led to credit rating downgrades and difficulties in discounting for some bills during circulation, the system would adjust the transmission probability of the bill paths involved in that region upward by 10%-15%. Conversely, if a certain type of bill exhibited greater stability due to policy support or an improved market environment, its transmission probability could be adjusted downward by 5%-10%. Furthermore, the system continuously learns new risk transmission patterns through machine learning models and automatically adjusts previously set initial values, thereby improving the timeliness and accuracy of probability assignments. After dynamic adjustments, the system generates adjusted risk transmission probabilities and further conducts a rationality assessment to ensure the adjusted probability values ​​are logically consistent and reasonably distributed. For example, if the transmission probability of a particular path increases from 0.65 to 0.89 after adjustment, the system verifies whether the path has indeed experienced a major risk event or market change. It also compares the trend of changes in the path nodes before and after the adjustment to confirm that the adjustment is well-founded, thereby forming a risk transmission probability assessment with reasonableness. Finally, based on the assessed risk transmission probability, the system constructs a risk transmission probability allocation for the bills. This allocation records the corresponding transmission probability value for each risk transmission path and categorizes and organizes them by path type, risk level, time period, and other dimensions. For example, in a batch of bills in the interbank bill discount market, the system identified a total of 1,200 assessed risk transmission paths. Of these, 35% were classified as "high-frequency trading" with an average transmission probability of 0.68; 42% were classified as "credit risk" with an average transmission probability of 0.74; and 23% were classified as "price anomaly" with an average transmission probability of 0.61.These paths and their corresponding transmission probability values ​​constitute a complete risk transmission probability distribution, which can be used in subsequent risk hotspot identification and strategy matching modules. In summary, by extracting risk transmission related factors from the assessed risk transmission paths and classifying them, and then completing the initial assignment, rationality check, dynamic adjustment and evaluation, a scientific and reasonable bill risk transmission probability distribution is finally constructed. The entire process realizes the refined modeling of the risk transmission intensity in the bill transaction process. This dynamic probability assignment method driven by multi-dimensional factors has been deployed and applied in the actual risk control systems of many commercial banks and bill trading platforms, effectively improving the accuracy of bill transaction risk identification and response efficiency, especially during periods of increased market volatility or regulatory policy adjustments. It can help institutions capture potential risk paths in a timely manner and optimize risk response measures.

[0035] In a specific embodiment, the identifying risk hotspots of the bill based on the risk transmission probability map to obtain the risk hotspot distribution includes: Dividing the risk transmission probability map into risk areas to obtain a set of risk areas, and marking risk characteristics of the set of risk areas to obtain risk areas marked with risk characteristics; Performing risk transmission link analysis on the bill based on the risk area with the risk characteristics marked to obtain a risk transmission link set, and performing link risk strength assessment on the risk transmission link set to obtain an assessed link risk strength; Performing preliminary risk hotspot identification on the bill based on the risk intensity of the assessment link to obtain preliminary risk hotspots, and performing hotspot stability analysis based on the preliminary risk hotspots to obtain stable risk hotspots; A risk hotspot distribution is constructed for the bill based on the stable risk hotspot to obtain a risk hotspot distribution.

[0036] Specifically, identifying risk hotspots for the bills based on the risk transmission probability map and generating a risk hotspot distribution is a key step in the overall bill transaction risk control method for locating and structurally representing high-risk areas. This step, by combining the spatial distribution characteristics of risk transmission paths, link strength assessment, and stability analysis, identifies areas with unusually high frequency clusters of risk events, nodes with peak risk transmission intensity, and risk concentration centers in different transaction scenarios within the complex bill transaction network. Ultimately, a risk hotspot distribution with both business interpretability and application value is constructed. The specific implementation process includes four core stages: first, the risk transmission probability map is divided into risk regions, forming a set of risk regions, and these regions are labeled with risk characteristics to generate risk regions with labeled risk characteristics; second, risk transmission link analysis is conducted based on these risk regions to extract a set of risk transmission links and assess their link risk strength to obtain the assessed link risk strength; third, based on these assessment results, risk hotspots are preliminarily identified, forming preliminary risk hotspots, and further hotspot stability analysis is conducted to screen for stable risk hotspots; fourth, based on this, a structured construction of the bill risk hotspot distribution is completed, resulting in the output of the risk hotspot distribution. In the first step, the system takes as input a risk transmission probability graph generated by a multi-stage risk transmission simulation. This graph uses nodes to represent the states of a bill at different transaction stages (e.g., "normal circulation," "high-frequency rediscount," "nearing redemption," and "high-risk transaction"), while edges represent the transmission paths between states and their corresponding transmission probabilities. For example, in the interbank bill discount market, the risk transmission probability graph for a bill contains 12 key state nodes and 45 transmission paths, with transmission probabilities on each path ranging from 0.1 to 0.9. The system performs cluster analysis on the spatial distribution characteristics of these nodes and paths, dividing nodes with similar risk transmission behavior into several risk regions, such as "early circulation," "mid-term trading," "pre-redemption risk exposure," and "default trigger," thereby forming a risk region cluster. The system then labels each risk region with risk characteristics, such as "high-frequency trading," "frequent credit rating declines," and "violent discount rate fluctuations," thereby generating risk regions with labeled risk characteristics. Next, the system conducts risk transmission chain analysis on the bill based on these labeled risk regions. The core of this analysis lies in identifying transmission links that span multiple risk areas and could potentially trigger large-scale risk contagion. For example, using graph theory's shortest path algorithms or key path mining techniques, the system can identify a complete chain of events that begins with "normal circulation," progresses through "high-frequency rediscount," "credit rating downgrades," "transfers nearing the due date," and ultimately reaches "default." These chains often represent the typical path by which risk gradually accumulates and ultimately erupts during bill transactions. The system aggregates all such paths into a risk transmission chain set and further assesses the risk strength of each chain.Evaluation criteria include the overall transmission probability of a link, the number of high-risk nodes involved in the link, and whether the link has been involved in historical default cases. For example, a link with a cumulative transmission probability of 0.83, including two "credit risk" nodes and one "price anomaly" node, and 17 similar paths leading to payment defaults in the past year, is assessed as a high-risk link with a link risk intensity score of 8.6 out of 10. The scores of all links constitute the assessed link risk intensity. The system then conducts a preliminary risk hotspot identification for the bill based on the assessed link risk intensity. "Risk hotspots" are nodes or regions in the risk transmission path that bear a high transmission responsibility, exhibit significant risk aggregation, or appear repeatedly in high-risk links. For example, in the example above, the "high-frequency rediscount" node appears in over 60% of the high-risk links, with an average transmission probability of 0.76, indicating its central role in the overall risk transmission network. The "credit rating downgrade" node is also identified as a potential hotspot due to its high correlation with subsequent default events (correlation coefficient of 0.81). The system labels these nodes as preliminary risk hotspots, forming a preliminary risk hotspot. To further improve identification accuracy, the system also conducts stability analysis on these preliminary hotspots, examining whether they consistently exhibit high-risk characteristics across different time windows and market environments. For example, if a node is identified as a risk hotspot in the first quarter of 2024, but its transmission probability drops significantly to below 0.35 in the second quarter, its risk performance is unstable and should be eliminated. Conversely, if its transmission probability remains above 0.70 for four consecutive quarters, it is confirmed as a stable risk hotspot. Through this process, the system ultimately identifies stable risk hotspots. Finally, based on these stable risk hotspots, the system constructs a structured distribution of bill risk hotspots. This construction process not only categorizes the hotspot nodes but also describes their distribution characteristics under different transaction scenarios. For example, during periods of tight liquidity, the stable risk hotspots for a certain type of bill are primarily concentrated in the "high-frequency rediscount" and "credit rating downgrade" nodes, accounting for 42% and 35%, respectively. During periods of market stability, the risk hotspots are more concentrated in the "transfer nearing the due date" node, accounting for 58%. Based on this, the system categorizes different types of risk hotspots by scenario and constructs a comprehensive risk hotspot distribution encompassing "areas with abnormally high frequency of risk events," "nodes with peak risk transmission intensity," and "risk concentration centers under different transaction scenarios." For example, in a certain batch of bills, 15 stable risk hotspot nodes were identified, including six "areas with abnormally high frequency of risk events," five "nodes with peak risk transmission intensity," and four "risk concentration centers under different transaction scenarios," forming a complete risk hotspot distribution.In summary, by dividing risk areas and annotating their characteristics from the risk transmission probability map, identifying risk transmission links and assessing their strength, preliminarily identifying and screening stable risk hotspots, and ultimately constructing a risk hotspot distribution with multi-dimensional characteristics, the entire process achieves accurate identification and structured representation of risk areas in the bill trading process. This method has been deployed in the actual risk control systems of multiple financial institutions, helping them identify hundreds of high-risk nodes and improving the targeted and forward-looking nature of bill trading risk prevention and control. Especially in the context of a complex and changing market environment and frequent regulatory policies, this method can effectively assist institutions in dynamically adjusting risk response strategies and ensuring the stable operation of the bill market.

[0037] In a specific embodiment, the risk transmission link analysis of the bill based on the risk area based on the risk feature annotated to obtain a risk transmission link set includes: Performing boundary feature extraction on the risk area with the risk feature marked to obtain a risk boundary feature set, and performing risk connectivity analysis on the risk boundary feature set to obtain a risk connectivity relationship matrix; Performing risk path topology analysis on the bill based on the risk connectivity relationship matrix to obtain a risk path topology map, and performing critical path segmentation on the risk path topology map to obtain a risk transmission sub-path set; Performing path feature fusion processing on the bill based on the risk conduction sub-path set to obtain a fused conduction feature sequence, and performing conduction strength calculation on the fused conduction feature sequence to obtain a path conduction strength graph; Based on the path conduction strength graph, link clustering analysis is performed on the bill to obtain an initial link category set, and the initial link category set is subjected to link optimization and reorganization to obtain a risk conduction link set, wherein the risk conduction link set includes a high-frequency conduction main link, a risk amplification branch link and a risk accumulation ring link.

[0038] Specifically, analyzing the risk transmission links of the bills based on the risk regions labeled with risk characteristics to generate a risk transmission link set is a key step in achieving structured identification and classification modeling of risk paths within the entire bill transaction risk control method. Starting from the risk regions that have been demarcated and labeled with risk characteristics, this step combines technical means such as boundary feature extraction, connectivity analysis, path topology construction, feature fusion processing, and link clustering optimization to ultimately generate a set of business-relevant and actionable risk transmission links. These links encompass various types, including high-frequency transmission main links, risk amplification branches, and risk accumulation loops, providing high-quality input support for subsequent risk hotspot identification and strategy matching. In specific implementation, the system first extracts boundary features from the risk regions labeled with risk characteristics. These risk regions have been defined and labeled with risk characteristics in a previous risk transmission probability map, such as "high-frequency trading concentration," "frequent credit rating declines," and "dramatic discount rate fluctuations." To further analyze the interactions between these regions, the system uses graph theory boundary detection algorithms or network node degree analysis methods to extract the connecting boundary features between each risk region, forming a risk boundary feature set. The system then conducts a risk connectivity analysis on these boundary features, calculating indicators such as the frequency of cross-regional bill transfers between different regions, the risk level overlap, and the density of risk transmission nodes. Based on this, it constructs a risk connectivity matrix. For example, in the interbank bill discount market, a certain bill undergoes cross-regional transfers between a "high-frequency rediscount zone" and a "credit rating downgrade zone" an average of 3.2 times per month, and the node density between these two zones reaches 18 per square kilometer (measured in logical space), indicating strong connectivity and the potential for risk interaction between the two. Next, the system conducts a risk path topology analysis on the bill based on this risk connectivity matrix, constructing a risk path topology map covering all risk zones. This map not only reflects the bill's flow paths between different risk zones but also depicts the hierarchical structure and dependencies between these paths. For example, the system identifies a typical path: "normal flow → high-frequency rediscount → credit rating downgrade → redemption default," and identifies multiple key nodes along this path, such as "first rediscount institution," "point at which credit rating drops below BBB," and "more than four transfers within 30 days before redemption." The system then performs critical path segmentation on this topology, breaking down the complex path into several sub-paths with specific functions and propagation characteristics, forming a set of risk transmission sub-paths. These sub-paths typically include: the **invoice endorsement chain** (i.e., the order in which an invoice circulates between multiple trading entities), the **funds settlement channel** (i.e., the path of payment funds corresponding to the invoice), and the **credit transmission chain** (i.e., the process by which credit risk gradually spreads between trading parties).For example, the system discovered that a bill endorsement chain involved six financial institutions, with multiple reverse transactions between the third and fifth, forming a typical credit risk transmission path. The system then performed path feature fusion based on this set of risk transmission subpaths, integrating key features from different subpaths to form a fused transmission feature sequence. This process is primarily accomplished through feature concatenation, weight assignment, and normalization, ensuring comparability and aggregation of data across different dimensions. For example, for the "bill endorsement chain" subpath, the system extracted features such as "number of endorsements," "number of participating institutions," and "distribution of institutional credit ratings." For the "fund settlement channel" subpath, it extracted metrics such as "settlement amount volatility," "settlement cycle," and "liquidity trend." And for the "credit transmission link" subpath, it focused on variables such as "credit rating change frequency," "default correlation coefficient," and "historical correlated risk events." After combining these features in a unified format, the system further calculated the transmission strength and generated a path transmission strength map. This map records information such as the transmission contribution of each node in each path, the path's overall carrying capacity, and the rate of risk accumulation along the entire path. For example, the highest node transmission contribution for a particular path is 0.87 (out of a maximum score of 1), indicating that this node significantly amplifies risk within the path. The overall path's carrying capacity is 120%, indicating that it can withstand risk pressures beyond its design limits. Furthermore, the path's risk accumulation rate reaches 5.3% per day, indicating a rapid risk transmission rate. Finally, the system performs link clustering analysis on the bills based on the aforementioned path transmission strength graph, identifying groups of paths with similar risk characteristics and forming an initial set of link categories. Clustering methods can employ K-means, DBSCAN, or a graph-embedding-based community discovery algorithm, automatically grouping based on multi-dimensional metrics such as transmission strength, path length, and node density. For example, the system identifies three typical link types: paths characterized by high-frequency transactions, denoted as "high-frequency transmission main links"; paths exhibiting significant risk amplification at certain nodes, denoted as "risk amplification branch links"; and paths characterized by circular transactions or repeated transfers, denoted as "risk accumulation loop links." To improve the accuracy and practicality of the identification results, the system further optimizes and reorganizes the initial set of link categories, eliminating redundant paths, merging similar paths, and correcting misjudged paths, ultimately generating a set of risk transmission links. For example, in a real-world application in an interbank bill discount market, the system identified 420 risk transmission paths in a batch of bills. After cluster analysis, these paths were divided into three categories: 138 high-frequency transmission main links, accounting for 32.9%; 192 risk amplification sub-links, accounting for 45.7%; and 90 risk accumulation loop links, accounting for 21.4%.Among them, the average transmission intensity of each path in the "risk amplification branch link" is 28% higher than that of other paths, and 47% of its nodes belong to trading entities with credit ratings below A, indicating a high level of risk exposure. Through this process, the system not only realizes the refined identification and classification of risk transmission paths in the bill transaction process, but also provides a solid data foundation and technical support for the subsequent identification of risk hotspots and the formulation of response strategies. In summary, by extracting boundary features and analyzing connectivity of risk areas with marked risk characteristics, constructing a risk path topology map and performing key path segmentation, and then combining path feature fusion and transmission intensity calculation, finally generating a set of risk transmission links through link clustering and optimized reorganization, the entire process realizes the comprehensive modeling and structured expression of the risk transmission path of bill transactions. This method has been implemented in the risk control systems of many financial institutions, effectively improving the accuracy of bill transaction risk identification and response efficiency, especially showing significant advantages in complex transaction structures and high-risk transaction scenarios.

[0039] In a specific embodiment, the preliminary risk hotspot identification of the bill based on the risk strength of the assessment link to obtain the preliminary risk hotspot includes: Performing risk intensity discretization processing on the risk intensity of the assessment link to obtain a risk intensity discrete value, and defining a discrete value range of the risk intensity discrete value to obtain a defined risk intensity discrete value; Constructing a risk hotspot pre-discrimination standard for the bill based on the defined risk intensity discrete value to obtain a risk hotspot pre-discrimination standard; performing a preliminary risk hotspot identification operation on the bill based on the risk hotspot pre-identification criteria to obtain a preliminary identified risk hotspot, and evaluating the credibility of the preliminary identified risk hotspot to obtain a preliminary identified risk hotspot with an evaluated credibility; Based on the preliminary identification of risk hotspots of the credibility, risk correlation factors of the bill are mined to obtain risk hotspot correlation factors, and hotspots of the bill are constructed based on the risk hotspot correlation factors to obtain preliminary risk hotspots.

[0040] Specifically, the preliminary risk hotspot identification of the bill based on the risk strength of the assessment link to obtain preliminary risk hotspots is one of the core sub-steps in the risk hotspot identification process. Its purpose is to extract nodes or areas with high risk transmission potential from the obtained link risk strength data, and to construct preliminary risk hotspots with business significance and operability through a series of structured processing methods. This process includes four key stages: first, the risk strength of the assessment link is discretized to form a discrete value of risk strength, and its value range is further defined to generate a defined discrete value of risk strength; second, based on the defined discrete value, a pre-discrimination standard for bill risk hotspots is constructed to form a pre-discrimination standard for risk hotspots, and its standard integrity is checked to ensure that the standard is comprehensive and logically consistent; third, based on the pre-discrimination standard, a preliminary identification operation of bill risk hotspots is carried out to generate a preliminary identification risk hotspot, and the credibility of the identification result is evaluated to ensure the identification quality; fourth, the correlation factors behind the preliminary identification results are further explored to form risk hotspot correlation factors, and finally a complete preliminary risk hotspot is constructed. In the first step, the system first discretizes the risk strength of the assessment link generated in the previous step. This assessment of link risk intensity typically consists of risk scores across multiple paths. For example, in the interbank bill discount market, 1,200 risk transmission paths for a particular type of bill were identified, each corresponding to a link risk intensity score ranging from 0.1 to 9.8 (out of a maximum of 10). To facilitate subsequent analysis and assessment, the system uses a discretization method to divide these continuous values ​​into several tiers, such as "low risk" (0.0–3.0), "medium risk" (3.1–6.0), "high risk" (6.1–8.5), and "extremely high risk" (8.6–10.0). These discrete risk intensity values ​​are then generated. The system then defines the range of these discrete values ​​to ensure that each tier is reasonable and effectively distinguishes different risk levels. For example, it confirms that paths within the "extremely high risk" interval have a high historical default rate or frequently appear in high-risk transaction chains, thus generating the defined discrete risk intensity values. The system then uses these defined discrete risk intensity values ​​to construct pre-deterministic criteria for identifying bill risk hotspots. The core function of this standard is to provide a unified basis for subsequent risk hotspot identification. For example, the system sets the following rules: If a node appears on more than three "very high risk" links and its average transmission probability is higher than 0.75, it is identified as a potential risk hotspot. If a node appears on at least five "high risk" links or higher and the historical default rate of its path exceeds 20%, it is also marked as a candidate risk hotspot. In addition, the system also considers the node's centrality indicators within the entire transmission graph (such as betweenness centrality and closeness centrality) as supplementary judgment criteria.Once constructed, the system performs a completeness check on the pre-identification criteria for risk hotspots to ensure that all risk levels have corresponding judgment logic and to avoid omissions or duplication of criteria. This ultimately results in the final, validated pre-identification criteria for risk hotspots. The system then performs preliminary identification of bills based on these pre-identified criteria. Specifically, the system traverses all bill paths and their corresponding nodes, matching those nodes that meet the criteria one by one according to pre-set criteria. For example, in a risk transmission probability graph for a particular bill, the "High-Frequency Rediscount" node appears on four "Extremely High Risk" links with an average transmission probability of 0.82. Three of the paths in which it resides have experienced actual defaults, making it a preliminary risk hotspot. The "Credit Rating Downgrade" node appears on six "High Risk" links with an average transmission probability of 0.76. The overall default rate of its path is 22%, also meeting the pre-set criteria, and is therefore included in the preliminary identification of risk hotspots. After identification, the system conducts a credibility assessment of the results, primarily through cross-validation, comparison with historical data, and expert review. For example, if a node meets the criteria but its path has a default record only at the historical average (approximately 12%), it may be excluded from the high-credibility hotspot, thus forming a preliminary identification of risk hotspots for credibility assessment. Finally, the system further conducts in-depth factor mining on these initially identified risk hotspots, analyzing the underlying risk drivers behind these hotspots to support subsequent stability analysis and strategy matching. For example, the system found that the "high-frequency rediscount" node became a hotspot primarily because the bills involved in it primarily involved non-standard transactions between small and medium-sized financial institutions, with a transaction frequency far exceeding the industry average (an average of 4.2 times per month). The "credit rating downgrade" node was closely associated with the presence of multiple counterparties with credit ratings below BBB-. Furthermore, the system identified that some hotspots were highly correlated with liquidity constraints during specific time periods (such as quarter-end and year-end), indicating that they were significantly affected by market conditions. By extracting these deep factors, the system forms risk hotspot correlation factors and optimizes and adjusts the preliminary identification results accordingly. For example, it eliminates nodes that are misjudged due to accidental factors and merges multiple nodes caused by similar mechanisms, ultimately constructing more representative and explanatory preliminary risk hotspots. In summary, by discretizing the risk intensity of the assessment link and defining its value range, constructing risk hotspot pre-identification standards and conducting preliminary identification, and then combining credibility assessment with deep factor mining, the system can scientifically and systematically identify key nodes with potential risk aggregation effects in the bill transaction process. This process has been deployed and applied in the actual risk control systems of many commercial banks and bill trading platforms, helping them to accurately locate hundreds of preliminary risk hotspots and significantly improve risk identification efficiency and prevention and control capabilities.Especially during periods of severe market volatility or regulatory policy adjustments, this method can respond quickly to changes and provide strong support for institutions to formulate targeted risk response strategies.

[0041] The above describes the transaction risk control method of the bill in the embodiment of the present invention. The following describes the transaction risk control device of the bill in the embodiment of the present invention. Figure 2 In one embodiment of the present invention, a device for controlling transaction risks of bills includes: A mining module 21 is configured to perform data mining on the historical transaction data of the bill using a preset data mining technology to obtain bill transaction behavior characteristics; A modeling module 22 is configured to perform state space modeling on the bill based on the bill transaction behavior characteristics to obtain a transaction state transition matrix; A simulation module 23 is configured to simulate a multi-stage risk transmission path of the bill based on the transaction state transition matrix to obtain a risk transmission probability map; Identification module 24, configured to identify risk hotspots for the bill based on the risk transmission probability map to obtain a risk hotspot distribution; wherein the risk hotspot distribution includes areas where the frequency of risk events is abnormally concentrated, risk transmission intensity peak nodes, and risk concentration centers under different transaction scenarios; The matching module 25 is used to match the risk response strategies of the bills based on the risk hotspot distribution to obtain a bill risk response strategy combination.

[0042] In this embodiment, for the specific implementation of each unit in the above device embodiment, please refer to the above method embodiment, which will not be repeated here.

[0043] Reference Figure 3 In an embodiment of the present invention, a computer device is also provided, wherein the internal structure of the computer device can be as follows: Figure 3 As shown. The computer device includes a processor, memory, display screen, input device, network interface and database connected via a system bus. The processor of the computer design is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store the corresponding data in this embodiment. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, the above method is implemented.

[0044] Those skilled in the art will understand that Figure 3The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present invention and does not constitute a limitation on the computer device to which the solution of the present invention is applied.

[0045] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which implements the above-described method when executed by a processor. It is understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0046] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing the relevant hardware using a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the above-described method embodiments. Any reference to memory, storage, database, or other media provided herein and used in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double-speed SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct RAMbus dynamic RAM (DRDRAM), and RAMbus dynamic RAM.

[0047] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, apparatus, article, or method comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, apparatus, article, or method. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, apparatus, article, or method comprising the element.

[0048] The above description is only a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the description and drawings of the present invention, or directly or indirectly applied in other related technical fields, is also included in the patent protection scope of the present invention.

Claims

1. A method for controlling transaction risks of bills, characterized in that: The following steps are involved: Performing data mining on the historical transaction data of the bill using a preset data mining technology to obtain bill transaction behavior characteristics; Performing state space modeling on the bill based on the bill transaction behavior characteristics to obtain a transaction state transfer matrix; Performing a multi-stage risk transmission path simulation on the bill based on the transaction state transition matrix to obtain a risk transmission probability graph; Based on the risk transmission probability map, risk hotspots of the bill are identified to obtain a risk hotspot distribution; wherein the risk hotspot distribution includes areas where the frequency of risk events is abnormally concentrated, risk transmission intensity peak nodes, and risk concentration centers under different transaction scenarios; Based on the risk hotspot distribution, risk response strategies are matched for the bills to obtain a bill risk response strategy combination.

2. The method for controlling bill transaction risks according to claim 1, characterized in that: The data mining of the historical transaction data of the bill is performed using a preset data mining technology to obtain bill transaction behavior characteristics, including: Performing transaction behavior structure analysis on historical transaction data of the bill to obtain transaction behavior structure elements, and analyzing structural relationships of the transaction behavior structure elements to obtain transaction behavior structure relationships; Based on the transaction behavior structural relationship, a transaction pattern is constructed for the historical transaction data of the bill to obtain a transaction pattern set, and transaction behavior features in the transaction pattern set are extracted to obtain bill transaction behavior features.

3. The method for controlling bill transaction risks according to claim 1, characterized in that: The state space modeling of the bill based on the bill transaction behavior characteristics to obtain a transaction state transfer matrix includes: Defining state variables for the bill transaction behavior characteristics to obtain bill transaction state variables, and defining a state space range for the bill transaction state variables; Determining a state transition rule for the bill based on the state space range to obtain a bill state transition rule, and setting an initial state of the bill according to the bill state transition rule to obtain an initial transaction state; Simulating a state transition process of the bill based on the initial transaction state to obtain a state transition process, and calculating a state transition probability of the state transition process to obtain a state transition probability; A transaction state transfer matrix is ​​constructed for the bill based on the state transfer probability to obtain a transaction state transfer matrix.

4. The method for controlling bill transaction risks according to claim 1, characterized in that: The multi-stage risk transmission path simulation of the bill based on the transaction state transition matrix to obtain a risk transmission probability map includes: Dividing the transaction state transfer matrix into risk transmission stages to obtain a multi-stage risk transmission framework, and defining nodes in the multi-stage risk transmission framework to obtain a risk transmission framework with node definitions; Preliminarily setting the risk transmission path of the bill based on the risk transmission framework with node definitions to obtain an initial risk transmission path, and performing a path possibility assessment based on the initial risk transmission path to obtain an assessed risk transmission path; Allocating risk transmission probabilities for the bills based on the assessed risk transmission paths to obtain a risk transmission probability distribution, and adjusting transmission probabilities based on the risk transmission probability distribution to obtain an adjusted risk transmission probability distribution; A risk transmission probability map is constructed for the bill based on the adjusted risk transmission probability distribution to obtain a risk transmission probability map.

5. The method for controlling bill transaction risks according to claim 4, characterized in that: The step of allocating risk transmission probabilities for the bills based on the assessed risk transmission paths to obtain risk transmission probability allocations includes: Analyzing the risk transmission correlation factors of the assessed risk transmission path to obtain risk transmission correlation factors, and classifying the risk transmission correlation factors by factor correlation to obtain classified risk transmission correlation factors; Performing an initial assignment of risk transmission probability to the bill based on the classified risk transmission correlation factors to obtain an initial value of the risk transmission probability, and checking the rationality of the initial value of the risk transmission probability to obtain a checked initial value of the risk transmission probability; Dynamically adjusting the risk transmission probability of the bill based on the initial value of the risk transmission probability after the inspection to obtain a dynamically adjusted risk transmission probability, and performing an adjustment rationality assessment on the dynamically adjusted risk transmission probability to obtain a risk transmission probability with an assessed rationality; Based on the risk transmission probability of the rationality of the assessment, a risk transmission probability distribution is constructed for the bill to obtain a risk transmission probability distribution.

6. The method for controlling bill transaction risks according to claim 1, characterized in that: The step of identifying risk hotspots of the bill based on the risk transmission probability map to obtain a risk hotspot distribution includes: Dividing the risk transmission probability map into risk areas to obtain a set of risk areas, and marking risk characteristics of the set of risk areas to obtain risk areas marked with risk characteristics; Performing risk transmission link analysis on the bill based on the risk area with the risk characteristics marked to obtain a risk transmission link set, and performing link risk strength assessment on the risk transmission link set to obtain an assessed link risk strength; Performing preliminary risk hotspot identification on the bill based on the risk intensity of the assessment link to obtain preliminary risk hotspots, and performing hotspot stability analysis based on the preliminary risk hotspots to obtain stable risk hotspots; A risk hotspot distribution is constructed for the bill based on the stable risk hotspot to obtain a risk hotspot distribution.

7. The method for controlling bill transaction risks according to claim 6, characterized in that: The step of performing preliminary risk hotspot identification on the bill based on the risk intensity of the assessment link to obtain preliminary risk hotspots includes: Performing risk intensity discretization processing on the risk intensity of the assessment link to obtain a risk intensity discrete value, and defining a discrete value range of the risk intensity discrete value to obtain a defined risk intensity discrete value; Constructing a risk hotspot pre-discrimination standard for the bill based on the defined risk intensity discrete value to obtain a risk hotspot pre-discrimination standard; performing a preliminary risk hotspot identification operation on the bill based on the risk hotspot pre-identification criteria to obtain a preliminary identified risk hotspot, and evaluating the credibility of the preliminary identified risk hotspot to obtain a preliminary identified risk hotspot with an evaluated credibility; Based on the preliminary identification of risk hotspots of the credibility, risk correlation factors of the bill are mined to obtain risk hotspot correlation factors, and hotspots of the bill are constructed based on the risk hotspot correlation factors to obtain preliminary risk hotspots.

8. A bill transaction risk control device, characterized in that: include: A mining module, configured to perform data mining on the historical transaction data of the bill using a preset data mining technology to obtain bill transaction behavior characteristics; A modeling module, configured to perform state space modeling on the bill based on the bill transaction behavior characteristics to obtain a transaction state transfer matrix; a simulation module, configured to simulate a multi-stage risk transmission path of the bill based on the transaction state transition matrix to obtain a risk transmission probability graph; an identification module for identifying risk hotspots for the bill based on the risk transmission probability map to obtain a risk hotspot distribution; wherein the risk hotspot distribution includes areas where the frequency of risk events is abnormally concentrated, risk transmission intensity peak nodes, and risk concentration centers under different transaction scenarios; A matching module is used to match the risk response strategy of the bill based on the risk hotspot distribution to obtain a bill risk response strategy combination.

9. A computer device comprising a memory and a processor, wherein a computer program is stored in the memory, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images