Internet-based logistics supply chain financial risk control method

Abstract

The invention relates to the technical field of financial risk intelligent control, particularly discloses a logistics supply chain financial risk control method based on the Internet, and solves the fuzzification problem of supply chain financial risk dispersion and risk transfer through a risk early warning mechanism. The method comprises the following steps: S1, accessing multi-source data of a platform in real time through the Internet, and constructing a supply chain real-time data view according to a timestamp alignment algorithm; s2, running a preset risk early warning model based on the supply chain real-time data view, dynamically monitoring and analyzing to obtain a risk factor, and generating a risk early warning event when judging that the risk factor meets a preset early warning threshold value; s3, carrying out transferability evaluation analysis on the risk early warning event, judging whether risk dispersion or transfer can be carried out through a financial tool, if so, entering step S4, and if not, entering step S3; if not, continuing monitoring; and S4, automatically matching a marketization risk distribution tool adaptive to the type, the target and the scale of the risk early warning event from a pre-connected financial product interface library.

Description

Technical Field

[0001] This invention relates to the field of intelligent financial risk control technology, specifically to an internet-based method for controlling financial risks in logistics supply chains. Background Technology

[0002] Logistics supply chain finance involves logistics companies playing a leading role in supply chain financial services. It leverages data on logistics, information flow, and capital flow within the supply chain to provide financial services based on credit enhancement and financing to upstream and downstream enterprises. However, this model is characterized by multi-source, dynamic, and contagious risks, making traditional risk control methods insufficient to cope with the increasingly complex market environment.

[0003] Supply chain finance can help logistics companies optimize risk management. On the one hand, by monitoring the flow of funds, logistics, and information in the supply chain in real time, supply chain finance can promptly identify potential risks. On the other hand, supply chain finance can use financing insurance or credit guarantees to diversify and transfer risks. However, these traditional methods all rely on market-based means to redistribute risks, making it difficult to directly and effectively identify risk problems in a timely manner and to quickly and effectively achieve the process of risk diversification and transfer. Summary of the Invention

[0004] The purpose of this invention is to provide an internet-based logistics supply chain finance risk control method, which solves the technical problem of how to achieve the ambiguity of risk diversification and risk transfer in supply chain finance through a risk early warning mechanism.

[0005] The objective of this invention can be achieved through the following technical solutions: Internet-based logistics supply chain finance risk control methods include: S1. Real-time access and integration of multi-source data from logistics operation platforms, supply chain transaction platforms, market data sources, and enterprise credit platforms via the Internet, and construction of a real-time supply chain data view according to the timestamp alignment algorithm; S2. Based on the real-time data view of the supply chain, run the preset risk warning model, dynamically monitor and analyze to obtain risk factors, and generate a risk warning event when the risk factors meet the preset warning threshold; the risk factors include market risk, credit risk and operational risk. S3. Conduct a transferability assessment and analysis of risk warning events to determine whether the risk can be diversified or transferred through financial instruments: If so, proceed to step S4; If not, continue monitoring; S4. Automatically match market-based risk allocation tools from the pre-connected financial product interface library that are appropriate for the type, target, and scale of the risk warning event; S5. By calling smart contracts, automatically execute subscription, underwriting, or hedging operation instructions associated with the selected risk allocation tool to complete the automated diversification or transfer of risk, and bind the operation record with the original financing business for evidence storage.

[0006] Preferably, the method for constructing the real-time supply chain data view in step S1 is as follows: By using data association algorithms to match logistics waybill entities, supply chain transaction order entities, market transaction entities, and enterprise credit data entities, a real-time supply chain data view integrating the three flows—capital flow, information flow, and logistics—is formed. The three flows include logistics data, capital flow data, and information flow data. The method for obtaining this integrated supply chain real-time data view is as follows: Through formula Calculate to obtain entity and entity Attribute similarity ;entity and entity It can be any two of the following: logistics waybill entity, supply chain transaction order entity, market transaction entity, and enterprise credit data entity; in, The attribute type that participates in the matching; For the first Weight of class attributes; For entities and entity exist Similarity based on class attributes; and , For entities and entity exist Distance on class attribute Maximum value For entities and entity exist Distance on class attribute Minimum value; edit distance is used when the attribute is text type, Euclidean distance is used when the attribute is numeric type, and direct distance is used when the attribute is encoded for an exact match. ,otherwise ; Attribute similarity Matching threshold Compare and judge: like ≥ When, the entity is determined. and entity If a match is successful, the data is merged into related data under the same supply chain node to identify the successfully matched entities. Otherwise, it is determined to be a non-related entity.

[0007] Preferably, the risk monitoring and analysis method in step S2 includes: S21. Input the dynamic data in the real-time supply chain data view into the risk warning model according to the preset cycle; S22. Extract features from the data input to the risk warning model, including market risk correlation features, credit risk correlation features, and operational risk correlation features; S23. Use the LSTM neural network model to predict the trend of the extracted time-series feature data, and use the random forest algorithm to rank the risk factors of the extracted non-time-series features. S24. Calculate the real-time risk value of each risk factor based on the feature weights to form a dynamic monitoring sequence of risk factors.

[0008] Preferably, the real-time risk value of the risk factor is calculated as follows: Through formula Calculate the real-time risk value of the current risk factor. ; in, Risk factors The feature dimensions included For the first The weights of each feature; Risk factors The The normalized values ​​of each feature, and , Risk factors The original eigenvalues, This is the minimum value of the original eigenvalue. This is the maximum value of the original eigenvalue.

[0009] Preferably, the calculation method for determining whether a risk factor meets a preset early warning threshold is as follows: For any risk factor Its warning threshold The calculation method is as follows:

[0010] in, For this risk factor The mean in historical data, For this risk factor Risk sensitivity coefficient; Real-time risk value of risk factors With warning threshold Compare: when ≥ When the risk warning threshold is met, a risk warning event is generated.

[0011] Preferably, the method for assessing and analyzing transferability is as follows: S31. Screen risk factors with transferable characteristics from real-time data of the integrated supply chain; S32. Extract the transferability features from the screened risk factors, calculate them, and output the transferability index; S33. Construct a binary classification judgment model by combining the early warning threshold and transferability index of risk factors, and output the risk transferability judgment result; S34. Track the effects of transferred risks and update model parameters through a feedback mechanism to optimize the model.

[0012] Preferably, the method for obtaining the transferability index in step S32 is as follows: Through formula Calculate the transferability index ; in, As a transferable feature dimension, and ∈ ,and ≥3; These are the characteristic values ​​of the transferable risk factor; This is the normalized value of the eigenvalues ​​of the transferable risk factor; For the first Preset weights for each feature.

[0013] Preferably, the output method for the risk transferability judgment result in step S33 is as follows: when ≥ At that time, after a risk event warning is generated, the transferability index is performed. preset threshold range Perform a comparison, and 0 < < And determine: like ≥ If the risk is high, it is considered highly transferable, and the company is advised to transfer all risks. like < < If it is determined to be of medium transferability, it is recommended that the company diversify its risks. like ≤ If the transferability is low, it is recommended that the company bear the risk itself.

[0014] The beneficial effects of this invention are: (1) The real-time data view of the supply chain panorama is determined by real-time data fusion in step S1; and the intelligent risk dynamic early warning mechanism in step S2 enables timely scanning and quantification of risk information from the supply chain panorama data in a proactive and continuous manner, so as to ensure that the risk warning signal can be clearly identified in the early stage of risk exposure, thereby providing accurate trigger signals and disposal objects for subsequent risk transfer and handling.

[0015] (2) Through the construction of the transferability assessment model in step S3 and the intelligent diagnosis results of the early warning risks, the risk categories and scale with obvious market transfer potential are automatically identified, avoiding the blindness of risk disposal under the traditional method, and accurately guiding the limited risk control resources to risks that can be effectively resolved through financial instruments, thereby improving the overall efficiency and cost-effectiveness of risk disposal.

[0016] (3) Steps S4 and S5 enable automated matching and execution of the entire process, achieving seamless connection and automatic execution from "risk warning events" to "market-based financial instruments." This directly transforms the technological achievements of the risk warning mechanism into executable market operations, significantly shortening risk response time, reducing operational risks and execution costs, and truly realizing the marketization, efficient dispersion, and transfer of risks. Furthermore, smart contract notarization and information binding technologies ensure that risk warning information is accurately and completely recorded, providing a technical foundation for subsequent data backtracking and model optimization.

[0017] Of course, any product implementing this invention does not necessarily need to achieve all the advantages described above at the same time. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the steps of the Internet-based logistics supply chain finance risk control method of the present invention. Figure 2 This is a flowchart illustrating the risk monitoring and analysis method in step S2 of the present invention. Figure 3 This is a diagram illustrating the transferability assessment and analysis steps in step S3 of the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Please see Figure 1 As shown, this invention is an internet-based logistics supply chain finance risk control method, the method comprising: S1. Real-time access and integration of multi-source data from logistics operation platforms, supply chain transaction platforms, market data sources, and enterprise credit platforms via the Internet, and construction of a real-time supply chain data view according to the timestamp alignment algorithm; S2. Based on the real-time data view of the supply chain, run the preset risk warning model, dynamically monitor and analyze to obtain risk factors, and generate a risk warning event when the risk factors meet the preset warning threshold; the risk factors include market risk, credit risk and operational risk. S3. Conduct a transferability assessment and analysis of risk warning events to determine whether the risk can be diversified or transferred through financial instruments: If so, proceed to step S4; If not, continue monitoring; S4. Automatically match market-based risk allocation tools from the pre-connected financial product interface library that are appropriate for the type, target, and scale of the risk warning event; S5. By calling smart contracts, automatically execute subscription, underwriting, or hedging operation instructions associated with the selected risk allocation tool to complete the automated diversification or transfer of risk, and bind the operation record with the original financing business for evidence storage.

[0022] The above technical solution utilizes an internet-based logistics supply chain finance risk control method to construct a risk management system capable of intelligent risk early warning and risk transferability assessment, automating the execution of financial instruments. The execution entity is a risk control platform deployed on a server cluster. This transforms traditional post-event static risk management into dynamic early warning driven by real-time data fusion, automatically triggering subsequent market-based risk allocation processes based on the early warning results. The specific steps are as follows: First, step S1 utilizes the platform through API interfaces, data buses, and other technologies to access multi-source, heterogeneous streaming data in real time, including logistics tracking information, supply chain electronic orders, commodity market data, and enterprise credit snapshots. The platform includes a logistics operation platform, a supply chain transaction platform, market data sources, and an enterprise credit platform. By constructing a data fusion engine with timestamp alignment and entity parsing capabilities, and cleaning, aligning, and correlating this data, a time-consistent, real-time panoramic data view of the supply chain is generated. This view can serve as the sole and reliable data source for all subsequent dynamic risk analyses. Step S1 provides a real-time and integrated data foundation for risk warning. Then, step S2 uses data from the real-time data view of the supply chain to pre-train the preset risk warning model. By scanning and analyzing the real-time data view, and by quantifying the three dimensions of market risk, credit risk, and operational risk, the risk factors for dynamic monitoring and analysis of the risk warning model are obtained. Once the calculated value or comprehensive evaluation value of any risk factor exceeds the warning threshold dynamically calculated based on historical behavior, the system automatically generates a corresponding risk warning event. This risk warning event indicates that the system has proactively discovered potential risk points through the dynamic monitoring and analysis of the model, thus providing a practical basis for solving the problem of lagging risk perception and inability to proactively discover risks in traditional methods. The reason for including market risk, credit risk, and operational risk as risk factors is that logistics finance is an important path for the transformation and upgrading of logistics enterprises, and its risk prevention and control needs to take into account both business innovation and risk transmission characteristics. Logistics finance risks stem from information asymmetry, external environmental fluctuations, and operational process defects. Market risk, credit risk, and operational risk are intertwined and can form a risk resonance effect. Building a collaborative prevention and control system can effectively mitigate risks.

[0023] Next, step S3 uses transferability assessment analysis to analyze the risk attributes of risk warning events and outputs a quantifiable indicator, namely the transferability index. The purpose is to screen and classify the warning risks, and accurately identify the information of risks that are suitable for market transfer and diversification through financial instruments. This provides a basis for the next step of accurate matching and avoids ineffective or high-cost transfer operations.

[0024] Furthermore, step S4 involves matching risk allocation tools to step S3. After the risk in step S3 is assessed as transferable, the process searches and matches a threshold and a standardized product interface library. By confirming the characteristics such as the type, indicators, and scale carried in the risk warning event, the process finds market-based risk allocation tools such as futures contracts, insurance products, and credit derivatives that are appropriate to the threshold. This completes the automated mapping process and realizes the interface layer for risk allocation. Finally, step S5 calls the smart contract to automatically send corresponding instructions to the back-end system of the financial market or insurance institution by binding the matching tool. After the instruction is received and successfully executed, the entire process from the early warning event to the financial operation is completely recorded and encrypted to ensure the immutability of the entire risk transfer chain and maintain auditability. Specifically, the operation is to call the smart contract to automatically execute the subscription, underwriting or hedging operation instructions associated with the selected risk allocation tool, complete the automatic diversification or transfer of risk, and bind the operation record with the original financing business for evidence storage, thus completing the automatic diversification or transfer of risk and full-process traceability record.

[0025] As one embodiment of the present invention, the method for constructing the real-time supply chain data view in step S1 is as follows: By using data association algorithms, entities such as logistics waybills, supply chain transaction orders, market transactions, and enterprise credit data are matched to form a unified real-time supply chain data view encompassing the three flows: cash flow, information flow, and logistics. The unified supply chain data view is obtained as follows: Through formula Calculate to obtain entity and entity Attribute similarity ;entity and entity It can be any two of the following: logistics waybill entity, supply chain transaction order entity, market transaction entity, and enterprise credit data entity; in, The attribute type that participates in the matching; For the first Weight of class attributes; For entities and entity exist Similarity based on class attributes; and , For entities and entity exist Distance on class attribute Maximum value For entities and entity exist Distance on class attribute Minimum value; edit distance is used when the attribute is text type, Euclidean distance is used when the attribute is numeric type, and direct distance is used when the attribute is encoded for an exact match. ,otherwise ; Attribute similarity Matching threshold Compare and judge: like ≥ When, the entity is determined. and entity If a match is successful, the data is merged into related data under the same supply chain node to identify the successfully matched entities. Otherwise, it is determined to be a non-related entity.

[0026] In the aforementioned technical solution, a weighted multi-attribute entity matching algorithm is used to solve the problem of multi-source data association, laying the foundation for accurate risk profiling and ensuring that the entity data used for early warning is highly relevant and authentic. By acquiring different entities, specifically including logistics waybill entities (logistics platform), supply chain transaction order entities (transaction platform), market transaction entities (market data source), and enterprise credit data entities (enterprise credit platform), these four entities are merged to form a unified panoramic view encompassing logistics status, transaction background, seller credit, and market price of goods. Specifically, information integration is achieved by calculating the similarity between different entities, using the following formula: Confirm the acquisition of the entity and entity Attribute similarity ; Utilizing a defined entity and entity exist By using class attribute similarity and introducing a weighted multi-attribute entity matching algorithm, the problem of "information silos" caused by independent systems and inconsistent data standards is effectively overcome. This transforms the complex problem of multi-source data association into a quantifiable and calculable process, reduces the workload of manual mapping and cleaning, improves the efficiency and scalability of building real-time data views, and lays a high-quality data foundation for the accurate analysis of subsequent risk models.

[0027] As one embodiment of the present invention, please refer to Figure 2 As shown, the risk monitoring and analysis methods in step S2 include: S21. Input the dynamic data in the real-time supply chain data view into the risk warning model according to the preset cycle; S22. Extract features from the data input to the risk warning model, including market risk correlation features, credit risk correlation features, and operational risk correlation features; S23. Use the LSTM neural network model to predict the trend of the extracted time-series feature data, and use the random forest algorithm to rank the risk factors of the extracted non-time-series features. S24. Calculate the real-time risk value of each risk factor based on the feature weights to form a dynamic monitoring sequence of risk factors.

[0028] In the above technical solution, firstly, step S21 involves inputting snapshots of the real-time supply chain data view into the model at fixed intervals; then, step S22 performs feature extraction, specifically extracting the following features: for market risk, extracting "the volatility of the underlying commodity price over the past 7 days" and "changes in related commodity indices"; for credit risk, extracting "recent sentiment scores of the financing party" and "changes in the guarantee capacity of related parties"; for operational risk, extracting "average delay time in transit" and "number of abnormal inventory turnovers of warehouse receipt collateral"; next, step S23 inputs time-series features such as price volatility into an LSTM neural network to predict price trends over a specified future time period, such as 24 hours; for non-time-series or static features such as "sentiment scores" and "delay time," inputs... The random forest model outputs a ranking of the importance of each feature in predicting the "probability of default." This importance is ranked according to the delay public opinion score; the higher the score, the more severe the delay public opinion. For example, it was found that "deterioration of public opinion" has the highest importance in the current scenario. Finally, step S24 calculates the real-time risk value of each risk factor through feature weighting, forming a dynamic monitoring sequence of risk factors. In the above, the model obtained by combining time series prediction and random forest feature analysis performs dynamic feature extraction and risk value calculation for risk information, and adopts a dynamic threshold judgment method based on historical averages and fluctuations. This makes risk warnings not based on fixed rules, but on intelligent and dynamic calculations, significantly improving the accuracy and foresight of warnings and ensuring that the risks that trigger warnings are risks that truly need and are worth dealing with.

[0029] As one embodiment of the present invention, the real-time risk value of the risk factor is calculated as follows: Through formula Calculate the real-time risk value of the current risk factor. ; in, Risk factors The feature dimensions included For the first The weights of each feature; Risk factors The The normalized values ​​of each feature, and , Risk factors The original eigenvalues, This is the minimum value of the original eigenvalue. This is the maximum value of the original eigenvalue.

[0030] In the above technical solution, the real-time risk value of the risk factor is calculated using the formula. This technology enables data-driven analysis of credit risk factors, integrating characteristic indicators from different dimensions and sources into a comprehensive, priced, real-time risk value. Furthermore, this real-time risk value makes the severity of risk intuitive and measurable, providing a scientific basis for subsequent threshold comparisons and avoiding fuzzy judgments.

[0031] As one embodiment of the present invention, the calculation method for determining whether a risk factor meets a preset early warning threshold is as follows: For any risk factor Its warning threshold The calculation method is as follows:

[0032] in, For this risk factor The mean in historical data, For this risk factor Risk sensitivity coefficient; Real-time risk value of risk factors With warning threshold Compare: when ≥ When the risk warning threshold is met, a risk warning event is generated.

[0033] In the above technical solution, the historical average risk value within historical events and a preset risk sensitivity coefficient are obtained. The product of the risk factor and its standard deviation is used to dynamically determine the warning threshold. Average risk value within a historical time window This represents the baseline of the "normal" or "average" risk level for that factor; through risk factors Sample standard deviation of risk values ​​within a historical time window This quantifies the magnitude of the historical volatility of the risk factor. The risk sensitivity coefficient is a pre-configured constant greater than 0, and administrators can set it according to the importance of different risk factors and the business's risk appetite; for core risks that require strict monitoring (such as core buyer credit risk), a larger coefficient can be set. A higher value is set to increase the warning threshold, thereby reducing false alarms and only triggering an alarm when the risk increases significantly; for risks that require high sensitivity (such as collateral price risk), a lower value can be set. The value makes the warning threshold relatively low, thus enabling earlier and more sensitive detection of rising risks and reducing underreporting.

[0034] As one embodiment of the present invention, please refer to Figure 3 As shown, the method for assessing transferability is as follows: S31. Screen risk factors with transferable characteristics from real-time data of the integrated supply chain; S32. Extract the transferability features from the screened risk factors, calculate them, and output the transferability index; S33. Construct a binary classification judgment model by combining the early warning threshold and transferability index of risk factors, and output the risk transferability judgment result; S34. Track the effects of transferred risks and update model parameters through a feedback mechanism to optimize the model.

[0035] In the above technical solution, the transferability assessment of risk-related events, including the transferability assessment process of market, credit, and operational risks, specifically involves: Step S31, screening transferability characteristics such as "whether the risk has an active trading market," "whether the risk can be standardized and defined," and "whether historical risk data is sufficient (for pricing)"; Step S32, determining the transferability result by judging the calculated transferability index within a preset threshold range in Step S33, and further analyzing it in conjunction with the warning details from Step S2. For market risk (high transferability index value), it is recommended to transfer it entirely through financial instruments; for operational risk (low transferability index value), it is recommended that the enterprise bear or diversify it through internal management (such as changing carriers); finally, Step S34, tracking the effect of the transferred risks and updating model parameters through a feedback mechanism to optimize the model.

[0036] As one embodiment of the present invention, the method for obtaining the transferability index in step S32 is as follows: Through formula Calculate the transferability index ; in, As a transferable feature dimension, and ∈ ,and ≥3; These are the characteristic values ​​of the transferable risk factor; This is the normalized value of the eigenvalues ​​of the transferable risk factor; For the first Preset weights for each feature.

[0037] In the above technical solutions, the transferability index Through formula Calculation determined; transferability index This is a normalized scalar value, and a higher value indicates that the risk represented by the risk warning event is more suitable and easier to transfer and diversify through market-based financial instruments; among them, the transferability characteristic dimension... The total number of feature dimensions used to assess transferability includes five common dimensions: market liquidity, risk measurability, data availability, legal contract completeness, and risk correlation. Through the [missing information - likely a specific method or approach]... Preset weights for each feature This represents the importance of different characteristics in the judgment of transferability. For example, market liquidity can be set to the highest level because it is a prerequisite for whether risk can be transferred.

[0038] Through the transferability index The calculation can transform qualitative, expert-based judgments into quantitative model calculations. By measuring, normalizing, and weighting multiple key dimensions (M≥3), it ultimately outputs a clear numerical value. This reflects that, prior to the disposal of financial instruments, triggering numerous risk warnings allows for the use of transferability indices. To reasonably determine the degree and method of risk transfer and diversification of financial instruments.

[0039] As one embodiment of the present invention, the output method for the risk transferability judgment result in step S33 is as follows: when ≥ At that time, after a risk event warning is generated, the transferability index is performed. preset threshold range Perform a comparison, and 0 < < And determine: like ≥ If the risk is high, it is considered highly transferable, and the company is advised to transfer all risks. like < < If it is determined to be of medium transferability, it is recommended that the company diversify its risks. like ≤ If the transferability is low, it is recommended that the company bear the risk itself.

[0040] In the above technical solution, the determination is made through calculation. With threshold range By comparing the results, risk warning events can be automatically classified. For results with high transferability, i.e. ≥ It is recommended to implement full risk transfer (such as purchasing insurance or futures hedging); for moderate transferability ( < < It is recommended to diversify risk or retain some of it (e.g., purchase insurance with deductibles); for low transferability ( ≤ If the risk is not diversifying or transferring, then the company should assume the risk and strengthen its internal management.

[0041] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the embodiments of apparatus, devices, and non-volatile computer storage media are basically similar to the method embodiments, and therefore described more simply; relevant parts can be referred to the descriptions of the method embodiments.

[0042] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined in this application, they should all fall within the protection scope of the present invention.

Claims

1. A method for controlling logistics supply chain finance risks based on the Internet, characterized in that: The method includes: S1. Real-time access and integration of multi-source data from logistics operation platforms, supply chain transaction platforms, market data sources, and enterprise credit platforms via the Internet, and construction of a real-time supply chain data view according to the timestamp alignment algorithm; S2. Based on the real-time data view of the supply chain, run the preset risk warning model, dynamically monitor and analyze to obtain risk factors, and determine when the risk factors meet the preset warning threshold to generate a risk warning event; the risk factors include market risk, credit risk and operational risk; S3. Conduct a transferability assessment and analysis on the aforementioned risk warning events to determine whether the risk can be diversified or transferred through financial instruments: If so, proceed to step S4; If not, continue monitoring; S4. Automatically match market-based risk allocation tools from the pre-connected financial product interface library that are appropriate for the type, target, and scale of the risk warning event; S5. By calling smart contracts, automatically execute subscription, underwriting, or hedging operation instructions associated with the selected risk allocation tool to complete the automated diversification or transfer of risk, and bind the operation record with the original financing business for evidence storage.

2. The Internet-based logistics supply chain finance risk control method according to claim 1, characterized in that, The method for constructing the real-time supply chain data view in step S1 is as follows: By using a data association algorithm, entities such as logistics waybills, supply chain transaction orders, market transactions, and enterprise credit data are matched to form a unified real-time supply chain data view encompassing the flow of funds, information, and logistics. The three flows include logistics data, fund flow data, and information flow data. The method for obtaining this unified supply chain data view is as follows: Through formula Calculate to obtain entity and entity Attribute similarity The entity and entity It can be any two of the following: logistics waybill entity, supply chain transaction order entity, market transaction entity, and enterprise credit data entity; in, The attribute type that participates in the matching; For the first Weight of class attributes; For entities and entity exist Similarity based on class attributes; and , For entities and entity exist Distance on class attribute Maximum value For entities and entity exist Distance on class attribute Minimum value; edit distance is used when the attribute is text type, Euclidean distance is used when the attribute is numeric type, and direct distance is used when the attribute is encoded for an exact match. ,otherwise ; Attribute similarity Matching threshold Compare and judge: like ≥ When, the entity is determined. and entity If a match is successful, the data is merged into related data under the same supply chain node to identify the successfully matched entities. Otherwise, it is determined to be a non-related entity.

3. The internet-based logistics supply chain finance risk control method according to claim 1, characterized in that, The risk monitoring and analysis method in step S2 includes: S21. Input the dynamic data in the real-time supply chain data view into the risk warning model according to the preset cycle; S22. Extract features from the data input to the risk warning model, including market risk correlation features, credit risk correlation features, and operational risk correlation features; S23. Use the LSTM neural network model to predict the trend of the extracted time-series feature data, and use the random forest algorithm to rank the risk factors of the extracted non-time-series features. S24. Calculate the real-time risk value of each risk factor based on the feature weights to form a dynamic monitoring sequence of risk factors.

4. The internet-based logistics supply chain finance risk control method according to claim 3, characterized in that, The real-time risk value of the risk factor is calculated as follows: Through formula Calculate the real-time risk value of the current risk factor. ; in, Risk factors The feature dimensions included For the first The weights of each feature; Risk factors The The normalized values ​​of each feature, and , Risk factors The original eigenvalues, This is the minimum value of the original eigenvalue. This is the maximum value of the original eigenvalue.

5. The Internet-based logistics supply chain finance risk control method according to claim 4, characterized in that, The calculation method for determining whether a risk factor meets the preset early warning threshold is as follows: For any risk factor Its warning threshold The calculation method is as follows:

6. Among them, For this risk factor The mean in historical data, For this risk factor Risk sensitivity coefficient; Real-time risk value of risk factors With warning threshold Compare: when ≥ When the risk warning threshold is met, a risk warning event is generated.

7. The Internet-based logistics supply chain finance risk control method according to claim 2, characterized in that, The method for assessing and analyzing the transferability is as follows: S31. Screen risk factors with transferable characteristics from real-time data of the integrated supply chain; S32. Extract the transferability features from the screened risk factors, calculate them, and output the transferability index; S33. Construct a binary classification judgment model by combining the early warning threshold and transferability index of risk factors, and output the risk transferability judgment result; S34. Track the effects of transferred risks and update model parameters through a feedback mechanism to optimize the model.

8. The Internet-based logistics supply chain finance risk control method according to claim 6, characterized in that, The method for obtaining the transferability index in step S32 is as follows: Through formula Calculate the transferability index ; in, As a transferable feature dimension, and ∈ ,and ≥3; These are the characteristic values ​​of the transferable risk factor; This is the normalized value of the eigenvalues ​​of the transferable risk factor; For the first Preset weights for each feature.

9. The Internet-based logistics supply chain finance risk control method according to claim 7, characterized in that, The output method for the risk transferability judgment result in step S33 is as follows: when ≥ At that time, after a risk event warning is generated, the transferability index is performed. preset threshold range Perform a comparison, and 0 < < And determine: like ≥ If the risk is high, it is considered highly transferable, and the company is advised to transfer all risks. like < < If it is determined to be of medium transferability, it is recommended that the company diversify its risks. like ≤ If the transferability is low, it is recommended that the company bear the risk itself.

Images