A new energy commercial vehicle pre-loan risk assessment method

Abstract

The application discloses a new energy commercial vehicle pre-loan risk assessment method, relates to the cross field of financial technology and data processing, and can solve the credit risk problem caused by the disconnection between static declaration information and dynamic operation environment and the mismatch between asset value and operation scene. The specific technical scheme is as follows: operation declaration information of a financing applicant and real-time three-flow data corresponding to the information are obtained, the total score of the dynamic matching degree of the two is calculated, and the total score and the environmental impedance index are used to correct the basic credit score to obtain a dynamic declaration credibility score; a first cash flow simulation and a second cash flow simulation are performed in parallel, a first simulated cash flow based on the operation declaration information is generated, and a second simulated cash flow based on counterfactual parameters corrected by the real-time three-flow data is generated; finally, a differentiated credit strategy including a credit limit and an interest rate is generated according to the deviation degree between the simulated cash flows and in combination with the dynamic declaration credibility score. The application is used for new energy commercial vehicle pre-loan risk assessment.

Description

Technical Field

[0001] This invention relates to the intersection of financial technology and data processing, specifically to a method for pre-loan risk assessment of new energy commercial vehicles. Background Technology

[0002] With the rapid development of the new energy commercial vehicle financing lease market, existing risk control methods mainly rely on static credit data of borrowers, such as central bank credit information and bank statements, to predict the borrower's future operating status through historical data. However, the actual profitability of new energy commercial vehicles is highly dependent on real-time fluctuations in the freight market, the convenience of energy replenishment, and road conditions. Static assessment models are unable to verify the authenticity of operating routes or predict the cash flow for fulfillment in complex market environments, leading to prominent information asymmetry problems.

[0003] Currently, existing risk control models suffer from rigid consistency flaws, treating any deviation between a user's actual behavior and their declared plan as a risk or fraud signal and deducting points accordingly. However, commercial vehicle operations have dynamic game-like characteristics. Drivers often proactively adjust routes or suspend operations due to severe weather, road congestion, or energy price fluctuations. Such rational risk-averse behavior is easily misjudged as abnormal operations under existing algorithms, leading to the wrong rejection of loans for high-quality customers.

[0004] Meanwhile, existing asset valuation methods suffer from a mismatch between the subject and the asset risk. Traditional models rely solely on linear depreciation based on vehicle age and mileage, neglecting the sensitivity of new energy commercial vehicle battery lifespan to operational scenarios. The battery degradation rate of the same model can differ by several times between heavy-duty transport in plains and temperate regions and in high-altitude, cold mountainous areas. Current technology lacks a mechanism to map specific operational scenarios to physical asset depreciation, failing to identify the risk of a vehicle's residual value falling below the outstanding principal under certain scenarios. This risk can easily induce borrowers to default on payments and abandon the vehicle, resulting in bad debt losses. Summary of the Invention

[0005] To address the credit risk issues arising from the disconnect between static application information and dynamic operating environment, and the mismatch between asset value and operating scenario in existing new energy commercial vehicle financing and leasing businesses, this invention provides a pre-loan risk assessment method for new energy commercial vehicles. The technical solution is as follows:

[0006] According to a first aspect of the present invention, a method for pre-loan risk assessment of new energy commercial vehicles is provided, the method comprising:

[0007] Obtain operational declaration information from financing applicants and collect real-time three-flow data that have a spatiotemporal correspondence with the operational declaration information; the real-time three-flow data includes real-time freight flow data, real-time vehicle flow data, and real-time energy flow data;

[0008] The dynamic matching score between the operational declaration information and the real-time three-stream data is calculated, and the basic credit score of the financing applicant is corrected based on the dynamic matching score and the environmental resistance index indicated by the real-time three-stream data to obtain the dynamic credibility score of the declaration; wherein, the environmental resistance index is used to quantify the severity of the current operating environment;

[0009] Based on the dynamic score of the declaration credibility, the first cash flow simulation and the second cash flow simulation are executed in parallel to generate the first simulated cash flow with the operational declaration information as input and the second simulated cash flow with the counterfactual parameters corrected by the real-time three-flow data as input, respectively.

[0010] Based on the degree of deviation between the first simulated cash flow and the second simulated cash flow, and in conjunction with the dynamic score of the declaration credibility, a differentiated credit granting strategy including credit limit and interest rate is generated.

[0011] This invention provides a pre-loan risk assessment method for new energy commercial vehicles. First, it obtains the operational declaration information of the financing applicant and collects real-time three-flow data that has a spatiotemporal correspondence with the operational declaration information. The real-time three-flow data includes real-time freight flow data, real-time vehicle flow data, and real-time energy flow data. Then, it calculates the total dynamic matching score between the operational declaration information and the real-time three-flow data. Based on the total dynamic matching score and the environmental resistance index indicated by the real-time three-flow data, it corrects the financing applicant's basic credit score to obtain a dynamic score for declaration credibility. The environmental resistance index is used to quantify the severity of the current operating environment. Next, based on the dynamic score for declaration credibility, it executes a first cash flow simulation and a second cash flow simulation in parallel, generating a first simulated cash flow with operational declaration information as input and a second simulated cash flow with counterfactual parameters corrected by the real-time three-flow data as input. Finally, based on the deviation between the first and second simulated cash flows and combined with the dynamic score for declaration credibility, it generates a differentiated credit granting strategy that includes credit limits and interest rates. This invention's method, by introducing real-time freight flow, vehicle flow, and energy flow data, achieves a shift from "static observation of people to dynamic observation of scenarios." It solves the credit risk problem in existing new energy commercial vehicle financing and leasing businesses caused by the disconnect between static declaration information and dynamic operating environment, and the mismatch between asset value and operating scenario. Under the premise of ensuring credit security, it can accurately identify rational risk-averse behaviors and significantly improve the robustness of the risk control model and the accuracy of credit decisions.

[0012] As a further aspect of the present invention: the collection of real-time three-stream data that has a spatiotemporal correspondence with the operational reporting information includes:

[0013] Determine the planned operation route and work area in the operation application information;

[0014] Using the planned operation route and work area as the spatial reference and the current time or the reporting plan cycle as the time reference, the real-time order density and freight rate within the planned operation route and work area are retrieved from the data interface as real-time freight flow data, the operation density and traffic efficiency of the same type of vehicles are retrieved as real-time vehicle flow data, and the idle rate and energy price of charging and swapping facilities are retrieved as real-time energy flow data.

[0015] The method of this invention collects three-stream data by setting a spatiotemporal benchmark, ensuring the consistency between real-time data and declaration information in the spatial and temporal dimensions, and providing a data foundation for subsequent matching degree calculation and authenticity verification.

[0016] As a further aspect of the present invention: the calculation of the total dynamic matching score between the operational declaration information and the real-time three-stream data includes:

[0017] The sub-matching scores for the freight flow dimension are determined based on the deviation between the declared freight rate and the real-time freight rate, as well as the deviation between the declared frequency and the real-time order density; wherein, the deviation is the ratio of the absolute difference between the declared value and the real-time value to the real-time value.

[0018] The sub-matching score of the traffic flow dimension is determined based on the spatial overlap ratio between the declared route and the mainstream trajectory of the same type of vehicle; wherein, the spatial overlap ratio is the ratio of the length of the overlapping road segment to the total length of the declared route.

[0019] The sub-item matching score of the energy flow dimension is determined based on the deviation between the declared energy replenishment time and the actual waiting time calculated based on the real-time status of the charging and swapping facilities.

[0020] Obtain the preset weight coefficients for each dimension, and sum the weighted scores of each sub-item matching score to obtain the total dynamic matching score.

[0021] The method of this invention calculates the matching scores of three dimensions—freight flow, vehicle flow, and energy—separately, and then performs a weighted summation, thereby achieving a multi-dimensional quantitative comparison between operational declaration information and real-time data from these three flows. This provides a quantifiable basis for subsequent credit assessment.

[0022] As a further aspect of the present invention: the step of correcting the basic credit score of the financing applicant based on the total dynamic matching score and the environmental impedance index indicated by the real-time three-stream data to obtain a dynamic score for application credibility includes:

[0023] The dynamic matching score is mapped to a preset matching level range to determine the corresponding initial correction factor; wherein, when the dynamic matching score is lower than the preset standard, the initial correction factor is negative.

[0024] When the total dynamic matching score is lower than the preset standard and the environmental impedance index exceeds the preset impedance threshold, it is determined that the deviation between the operational declaration information and the real-time three-stream data belongs to rational risk avoidance behavior. A rational compensation value is generated based on the environmental impedance index, and the rational compensation value is used to correct the initial correction factor to obtain the final correction factor, and the final correction factor is not lower than zero; otherwise, the initial correction factor is used as the final correction factor.

[0025] The final correction factor is added to the basic credit score to obtain the dynamic score of the declaration credibility.

[0026] The method of this invention determines the initial correction factor by mapping the total dynamic matching score to the matching level range. When the total dynamic matching score is lower than the preset standard and the environmental impedance index exceeds the preset impedance threshold, the deviation is determined to be a rational risk-avoidance behavior and a rational compensation value is generated for correction. This achieves dynamic adjustment of the basic credit score and avoids the wrong rejection of high-quality customers due to environmental factors.

[0027] As a further aspect of the present invention: the environmental impedance index is calculated based on the real-time three-flow data, and the environmental impedance index is positively correlated with the scarcity of goods, the degree of road congestion, and the degree of energy replenishment congestion.

[0028] Wherein, the scarcity of goods is the ratio of the baseline order density to the real-time order density; the road congestion is the ratio of the designed traffic speed to the real-time average vehicle speed; and the refueling congestion is the ratio of the real-time queuing time to the baseline tolerable time.

[0029] The environmental impedance index is a weighted sum of the scarcity of goods, the degree of road congestion, and the degree of energy replenishment congestion.

[0030] The method of this invention identifies rational risk-averse behavior by setting a preset impedance threshold, avoiding misjudgments caused by environmental factors such as severe weather, traffic congestion, or energy price fluctuations, and providing technical support for the accurate identification of high-quality customers.

[0031] As a further aspect of the present invention: the parallel execution of the first cash flow simulation and the second cash flow simulation includes:

[0032] Using the fare, operating frequency, and route energy consumption in the operational declaration information as input parameters, and combined with the vehicle fixed costs, the net cash flow during the financing cycle is extrapolated according to the time series to generate the first simulated cash flow.

[0033] Distortion parameters in the operational declaration information that differ from the real-time three-stream data by more than a preset threshold are identified. The corresponding parameters in the real-time three-stream data are used to replace the distortion parameters to obtain counterfactual parameters. Using the counterfactual parameters as input, combined with the vehicle fixed costs, the net cash flow within the financing cycle is extrapolated according to the same time series to generate the second simulated cash flow.

[0034] The method of this invention distinguishes between the application scenario and the real-world scenario through parallel simulation and comparative calculation, enabling the risk assessment server to test the future repayment ability of the financing applicant and provide a basis for subsequent judgment of the risk level of credit granting.

[0035] As a further aspect of the present invention: the step of generating a differentiated credit granting strategy, including credit limit and interest rate, based on the degree of deviation between the first simulated cash flow and the second simulated cash flow, and in conjunction with the dynamic score of the declared credibility, includes:

[0036] Calculate the sum of the absolute values ​​of the differences between the first simulated cash flow and the second simulated cash flow at each time point, and calculate the proportion of the sum of the absolute values ​​of the differences to the cumulative total of the second simulated cash flow to obtain the core deviation index;

[0037] If the core deviation index exceeds the preset deviation tolerance threshold, it is determined that the operational reporting information has significant distortion, and the second simulated cash flow is used as the credit assessment benchmark; otherwise, the fusion weight of the first simulated cash flow and the second simulated cash flow is determined according to the core deviation index, and the weighted fusion result is used as the credit assessment benchmark.

[0038] The credit limit and interest rate are determined based on the dynamic score of the declared credibility and the credit assessment benchmark.

[0039] The method of this invention calculates the core deviation index between the first simulated cash flow and the second simulated cash flow, and forces the adoption of the second simulated cash flow when the core deviation index exceeds the preset deviation tolerance threshold; otherwise, it performs weighted fusion based on the core deviation index to achieve differentiated determination of credit assessment benchmarks, providing a basis for the generation of credit limits and interest rates.

[0040] As a further aspect of the present invention: the step of determining the credit limit and interest rate based on the dynamic score of the declared credibility and the credit assessment benchmark includes:

[0041] The cash flow coverage ratio and cash flow volatility index are calculated based on the credit assessment benchmark. The cash flow coverage ratio is the ratio of the cumulative value of the credit assessment benchmark to the cumulative value of principal and interest owed. The cash flow volatility index is the ratio of the standard deviation to the mean of the credit assessment benchmark.

[0042] A three-dimensional comprehensive score is calculated by combining the dynamic score of the declaration credibility, the cash flow coverage ratio, and the cash flow fluctuation index.

[0043] The three-dimensional comprehensive score is mapped to a preset risk classification table to determine the main risk level and the volatility sub-level.

[0044] The basic credit limit range is determined based on the primary risk level, and the interest rate fluctuation ratio is determined based on the secondary volatility level. These factors are then combined to generate the differentiated credit strategy.

[0045] The method of this invention calculates the cash flow coverage ratio and cash flow volatility index, and calculates a three-dimensional comprehensive score by combining the dynamic score of the declaration credibility, the cash flow coverage ratio and the cash flow volatility index. Then, the three-dimensional comprehensive score is mapped to a risk classification table to determine the main risk level and the volatility sub-level, thereby realizing the differentiated determination of the credit limit range and the interest rate fluctuation ratio.

[0046] As a further aspect of the present invention: before generating the differentiated credit granting strategy, the method further includes:

[0047] Extract environmental characteristic factors from the operational declaration information and the real-time three-stream data. The environmental characteristic factors include at least ambient temperature difference, road condition slope, and load.

[0048] The scenario-based depreciation acceleration factor is calculated based on the environmental characteristic factors to predict the residual value of assets at the end of the financing cycle; wherein, the scenario-based depreciation acceleration factor is positively correlated with the superimposed effect of the environmental temperature difference, the road slope, and the load.

[0049] If the remaining value of the asset is lower than the preset lower limit of the basic credit line, the credit line circuit breaker mechanism is triggered, and the basic credit line range is adjusted downward based on the remaining value of the asset. The adjusted credit line range is used as a constraint to generate the differentiated credit strategy.

[0050] The method of this invention extracts environmental characteristic factors and calculates scenario-based depreciation acceleration coefficients to predict the residual value of assets at the end of the financing cycle. When the residual value of assets is lower than the preset lower limit of the basic credit line, a credit line circuit breaker mechanism is triggered to adjust the range of the basic credit line downward. This realizes the verification of asset value constraints and provides an upper limit constraint for the generation of differentiated credit strategies.

[0051] As a further aspect of the present invention, the method further includes:

[0052] The operation declaration information, the data digest of the real-time three-stream data, the first simulated cash flow, the second simulated cash flow, and the differentiated credit granting strategy are subjected to cryptographic hashing to generate a unique digital fingerprint;

[0053] The digital fingerprint and timestamp are written into the blockchain evidence storage system to solidify the data of the entire pre-loan assessment process.

[0054] The method of this invention generates a unique digital fingerprint by performing cryptographic hash operations on the data digests of operational declaration information, real-time three-stream data, the first simulated cash flow, the second simulated cash flow, and the differentiated credit granting strategy. The digital fingerprint and timestamp are then written into a blockchain storage system, realizing the solidification and storage of data throughout the entire pre-loan assessment process. This ensures the immutability and traceability of the assessment data throughout the entire process.

[0055] According to a second aspect of the present invention, a pre-loan risk assessment device for new energy commercial vehicles is provided. The pre-loan risk assessment device for new energy commercial vehicles includes a processor and a memory. The memory stores at least one computer instruction, which is loaded and executed by the processor to perform the steps in the pre-loan risk assessment method for new energy commercial vehicles described above.

[0056] According to a third aspect of the present invention, a computer-readable storage medium is provided, the storage medium storing at least one computer instruction, the instruction being loaded and executed by a processor to perform the steps performed in the pre-loan risk assessment method for new energy commercial vehicles as described in any of the preceding claims.

[0057] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Attached Figure Description

[0058] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0059] Figure 1 This is a flowchart of the pre-loan risk assessment method for new energy commercial vehicles provided in an embodiment of the present invention. Detailed Implementation

[0060] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention.

[0061] This invention provides a pre-loan risk assessment method for new energy commercial vehicles, primarily applied on the risk assessment server side. It aims to address credit risk issues arising from the disconnect between static application information and the dynamic operating environment, and the mismatch between asset value and operating scenarios in existing new energy commercial vehicle financing lease businesses. By introducing real-time freight flow, vehicle flow, and energy flow data (real-time three-flow data) reflecting the objective physical world, this embodiment constructs a closed-loop risk control system that includes dynamic authenticity verification, rational deviation verification, dual-track simulation stress testing, and an asset value circuit breaker mechanism.

[0062] In this embodiment, the risk assessment server can be a server cluster based on a cloud computing architecture, or an independent high-performance computing node deployed in a financial institution's private cloud. This server possesses the capability to process massive amounts of heterogeneous data concurrently and can interact with external data sources through various means, such as API interfaces, Internet of Things (IoT) gateways, and blockchain node interfaces. Figure 1 As shown, the pre-loan risk assessment method for new energy commercial vehicles includes the following steps:

[0063] Step 101: Obtain the operational declaration information of the financing applicant and collect real-time three-stream data that have a spatiotemporal correspondence with the operational declaration information;

[0064] In this embodiment, the real-time three-flow data includes real-time freight flow data, real-time vehicle flow data, and real-time energy flow data. Specifically, real-time freight flow data reflects the transportation status of goods, real-time vehicle flow data reflects the operating environment of vehicles, and real-time energy flow data reflects the energy replenishment status.

[0065] In practice, when a user (such as a new energy heavy truck driver or logistics company owner) submits a financial leasing application through a mobile app or web page, the risk assessment server first parses the application message. The core content of this message is the operational declaration information. This operational declaration information is the user's self-description of their future repayment ability, and it must include specific planned operating routes (e.g., from Datong, Shanxi to Tangshan, Hebei) and operating areas (e.g., Caofeidian Port Area of ​​Tangshan Port). In addition, the operational declaration information typically includes static planning data such as the vehicle model parameters to be purchased, estimated average monthly mileage, estimated one-way freight rate, and estimated monthly working days.

[0066] For example, to verify the authenticity of this self-reported information, the server cannot rely solely on static credit reports; it must incorporate more comprehensive and objective real-time data. Therefore, the server will immediately initiate a data access procedure to retrieve real-time three-stream data that is highly consistent with the aforementioned operational declaration information in terms of both time (current and historical data) and spatial (declared routes and regions) from interfaces of third-party data service providers, logistics platforms, vehicle networking platforms, and energy infrastructure operators.

[0067] In one embodiment, real-time three-stream data that has a spatiotemporal correspondence with the collected operational reporting information includes:

[0068] Determine the planned operation routes and work areas in the operation application information;

[0069] Using the planned operation routes and work areas as spatial benchmarks and the current time or the reporting plan cycle as time benchmarks, the system retrieves real-time order density and freight rates within the planned operation routes and work areas as real-time freight flow data, the operation density and traffic efficiency of similar vehicles as real-time vehicle flow data, and the idle rate and energy price of charging and battery swapping facilities as real-time energy flow data from the data interface.

[0070] In practical applications, the specific composition and acquisition logic of real-time three-stream data are as follows:

[0071] 1. Real-time freight flow data (reflecting cargo transportation status): The server retrieves capacity supply and demand data for the target route and region from the real-time data interface of leading digital freight platforms or logistics exchanges. Specifically, the server locks onto the user-declared operating area and queries the real-time order density (e.g., the number of pending orders per square kilometer) and real-time freight rates within that area at the current moment and over the past period. Optionally, real-time freight flow data also includes the distribution of cargo types (e.g., coal, steel, express delivery, etc.) and the average tonnage of the cargo. This data directly reflects the cargo supply situation on the user-declared route and whether the freight rate level is consistent with the declared content.

[0072] 2. Real-time traffic flow data (reflecting vehicle operating environment): The server connects to the commercial vehicle networking platform or the traffic management department's road condition data interface. Specifically, the server retrieves the real-time operating density of vehicles of the same type as the user's intended purchase vehicle (e.g., both are 6x4 battery-swapping tractors) on the user's declared route. For example, if the user declares the route to be highly profitable, but real-time data shows very few vehicles of the same type on the route, or extremely low traffic efficiency (average speed, congestion index), the system will mark it as abnormal. Traffic efficiency data can also reflect the complexity of road conditions, such as whether there are long-distance congestion or traffic control measures.

[0073] 3. Real-time Energy Flow Data (Reflecting Energy Supply Status): For new energy commercial vehicles, the convenience of refueling directly determines operational efficiency. Therefore, the server accesses data from the charging and battery swapping infrastructure operation platform. Specifically, the system scans charging piles or battery swapping stations along the user's declared route and within the operating area to obtain their idle rate (i.e., current available charging guns / total number of charging guns or available batteries / total reserves) and energy prices (real-time electricity price or hydrogen price). It's also important to note that the energy flow data should include queuing time prediction data, which is directly related to the vehicle's effective operating time. Through these steps, the server logically constructs a digital twin environment parallel to the user's declared information, laying the data foundation for subsequent matching degree calculations.

[0074] The method of this invention collects three-stream data by setting a spatiotemporal benchmark, ensuring the consistency between real-time data and declaration information in the spatial and temporal dimensions, and providing a data foundation for subsequent matching degree calculation and authenticity verification.

[0075] Step 102: Calculate the total dynamic matching score between the operational declaration information and the real-time three-stream data, and adjust the basic credit score of the financing applicant based on the total dynamic matching score and the environmental resistance index indicated by the real-time three-stream data to obtain the dynamic credibility score of the declaration; among which, the environmental resistance index is used to quantify the severity of the current operating environment.

[0076] In one embodiment, calculating the total dynamic matching score between operational reporting information and real-time three-stream data includes:

[0077] Based on the deviation between the declared freight rate and the real-time freight rate, as well as the deviation between the declared frequency and the real-time order density, the sub-item matching scores of the cargo flow dimension are determined; whereby the deviation is the ratio of the absolute difference between the declared value and the real-time value to the real-time value.

[0078] The traffic flow dimension matching score is determined based on the spatial overlap ratio between the declared route and the mainstream trajectory of similar vehicles; where the spatial overlap ratio is the ratio of the length of the overlapping road segment to the total length of the declared route.

[0079] The sub-matching score of the energy flow dimension is determined based on the deviation between the declared energy replenishment time and the actual waiting time estimated based on the real-time status of charging and swapping facilities.

[0080] Obtain the preset weight coefficients for each dimension, and sum the weighted scores of each sub-item to obtain the total dynamic matching score.

[0081] In this embodiment, the server calculates the item matching scores for each of the three dimensions:

[0082] 1. Freight Flow Dimension Matching Score: The server compares the user's estimated one-way freight rate with the market average freight rate in real-time freight flow data; it also compares the user's declared monthly operating frequency with the order density in real-time freight flow data. If the declared freight rate is much higher than the market price, or the declared frequency is much higher than the limit that the order density can support, then this sub-item score will be low.

[0083] 2. Traffic Flow Dimension Matching Score: The server compares the user's declared driving route with the mainstream trajectories of similar vehicles in real-time traffic flow data. If most similar vehicles avoid the route (possibly due to traffic restrictions or extremely poor road conditions), but the user insists on taking this route, the score for this sub-item will be low.

[0084] 3. Energy Flow Dimension Matching Score: The server calculates the user's reported energy replenishment cost and time. If the idle rate of charging piles along the route is consistently zero or the failure rate is extremely high, but the user reports extremely high operational efficiency, this is logically invalid, and this sub-item will receive a low score.

[0085] The specific calculation logic is as follows:

[0086] 1. Matching score based on cargo flow dimension ( ): Calculated based on the deviation rate between declared freight rates and frequencies and actual market data. The formula is:

[0087]

[0088] in, and These represent the estimated one-way freight rate and the market average freight rate from real-time freight flow data, respectively. and These represent the monthly operational frequency and the objectively executable frequency calculated based on order density, respectively; k1 and k2 are preset sensitivity coefficients, typically set to 0.5.

[0089] 2. Traffic flow dimension matching score ( ): Calculated based on trajectory overlap. The formula is:

[0090]

[0091] in, The total length of the declared driving route. This refers to the length of the section of the route that overlaps with the mainstream trajectory of the same type of vehicle.

[0092] 3. Energy flow dimension matching score ( ): Calculated based on the percentage of available refueling time. The formula is:

[0093]

[0094] in, This refers to the total duration of a single refueling cycle. The estimated waiting time is derived from the predicted data on the availability of charging facilities along the route and the queue length.

[0095] The server performs a weighted summation using preset weighting coefficients. For example, let the preset weighting coefficients be: , , And satisfy The dynamic matching degree total score is as follows. The calculation formula is as follows: The total dynamic matching score is equal to the matching score of the cargo flow dimension multiplied by the cargo flow weight coefficient, plus the matching score of the vehicle flow dimension multiplied by the vehicle flow weight coefficient, plus the matching score of the energy flow dimension multiplied by the energy flow weight coefficient.

[0096] Suppose user A declares the operation of a dedicated coal transport line from Datong, Shanxi to Tangshan, Hebei. The system pre-sets the weighting coefficients for freight flow, vehicle flow, and energy flow as follows: , , Based on system comparison and calculation: the current coal freight rate for this route is stable, and the freight flow dimension matching score is [score missing]. Score: This route is a mainstream route for heavy trucks, with a traffic flow matching score of S. vehicle = 85 points; however, some battery swapping stations along the route are under maintenance, resulting in low idle rates and a low energy flow matching score. energy = 60 points.

[0097] According to the formula, the total dynamic matching score S is calculated. match = 90 × 0.5 + 85 × 0.3 + 60 × 0.2 = 45 + 25.5 + 12 = 82.5 points, which falls within the high-match interval [80, 100].

[0098] The method of this invention calculates the matching scores of three dimensions—freight flow, vehicle flow, and energy—separately, and then performs a weighted summation, thereby achieving a multi-dimensional quantitative comparison between operational declaration information and real-time data from these three flows. This provides a quantifiable basis for subsequent credit assessment.

[0099] In one embodiment, the applicant's basic credit score is adjusted based on the dynamic matching score and the environmental resistance index indicated by real-time three-stream data to obtain a dynamic credit score for the application, including:

[0100] The dynamic matching score is mapped to a preset matching level range to determine the corresponding initial correction factor; when the dynamic matching score is lower than the preset standard, the initial correction factor is negative.

[0101] When the total dynamic matching score is lower than the preset standard and the environmental impedance index exceeds the preset impedance threshold, it is determined that the deviation between the operational declaration information and the real-time three-stream data is a rational risk avoidance behavior. A rational compensation value is generated based on the environmental impedance index, and the rational compensation value is used to correct the initial correction factor to obtain the final correction factor, and the final correction factor is not lower than zero; otherwise, the initial correction factor is used as the final correction factor.

[0102] The final correction factor is added to the basic credit score to obtain the dynamic score of the application credibility.

[0103] In actual use, the server will execute a dynamic correction process based on the basic credit score:

[0104] 1. Determine the basic credit score: The server first retrieves static data such as the user's central bank credit report and historical repayment records, and uses a traditional credit scoring card model (such as the standard application scoring card A card model based on logistic regression) to calculate a basic credit score, which reflects the user's historical background.

[0105] 2. Determine the interval level: The server will calculate the total dynamic matching score. Mapped to a preset numerical range level. For example, [80, 100] is a high matching range, [60, 80) is a medium matching range, and [0, 60) is a low matching range.

[0106] 3. Determine the correction factor: The system assigns a correction factor to each interval. Specifically:

[0107] When in a high-match range, the correction factor is positive (e.g., +5 points), which means that words and actions are consistent, the environment is favorable, and credit score is improved.

[0108] When the match is in the middle range, the correction factor may be zero.

[0109] It should also be noted that when the numerical range level indicator matches the preset standard (i.e., it is in the low matching range), the correction factor is initially set to a negative value (e.g., -10 points), which means that the declaration is false and there is fraud or high risk.

[0110] 4. Calculate the dynamic score for application credibility: The final dynamic score for application credibility. It equals the base credit score plus a correction factor.

[0111] In practical applications, mechanically deducting points for low-match scores may inadvertently harm high-quality customers. For example, a user declares they will take Route A, but Route A is suddenly hit by a blizzard, forcing the user to take Route B or for the route to be suspended. This deviation is rational and should not be considered credit risk. Therefore, this embodiment introduces a rational deviation verification step. When the match score is determined to enter the low-match range and the correction factor is set to a negative value, the server will not immediately deduct points, but will automatically trigger the rational deviation verification subroutine:

[0112] 1. Quantitative Environmental Impedance Index: The server needs to quantitatively assess the severity of the current environment. To this end, the server calculates the environmental impedance index based on real-time three-stream data. The higher the index, the more severe the external operating environment and the worse the profitability.

[0113] 2. Threshold Determination and Qualitative Characterization: The server will calculate the environmental impedance index. Compare with a preset impedance threshold.

[0114] Example: Setting an impedance threshold Weight A sudden blizzard hit Tangshan, causing local orders to plummet to 50% of normal levels. Icy roads caused average vehicle speeds to drop to 40% of normal levels. The waiting time for refueling is the same as usual. ).

[0115] At this point, the calculation yields... .

[0116] because The system determines that the user's shutdown at this time is a "rational risk-averse" behavior that seeks to avoid harm and seek benefits.

[0117] Scenario A (Malicious Loan Fraud): If the environmental impedance index is lower than the impedance threshold (indicating that the environment is good, there is cargo, the road is open, and the power is sufficient), but the user still deviates significantly from the declared plan (for example, instead of picking up cargo, the user drives the vehicle to a non-operational area), the system will determine that it is an irrational deviation or suspected fraud, and maintain the decision that the correction factor is negative.

[0118] Scenario B (Rational Risk Avoidance): If the environmental impedance index is higher than or equal to the impedance threshold (indicating extremely harsh conditions), the user's shutdown or rerouting is interpreted by the system as a business instinct to seek advantage and avoid harm. The system determines this mismatch to be a rational risk avoidance behavior.

[0119] 3. Generation and replenishment of rational compensation value: For scenario B, the server will generate a positive rational compensation value based on the magnitude of the environmental impedance index. The more severe the environment, the larger this compensation value.

[0120] Supplement to the calculation method of rational compensation value:

[0121] Rational compensation value The calculation formula can be:

[0122]

[0123] in This is the preset compensation amplification factor, for example, a value of 20.

[0124] like The original correction factor was -10, and the compensation value was calculated to be 2.0. Points. The corrected correction factor = -10 + 10 = 0 points;

[0125] In extremely harsh environments, the system maintains its basic credit score unchanged.

[0126] Finally, the rational compensation value is used to correct the originally negative correction factor. Specifically, the corrected correction factor equals the original negative correction factor plus the rational compensation value. The system sets a bottom-line constraint: the corrected correction factor value is not lower than zero. This means that for users who rationally avoid risk, the system at least guarantees no deduction of points, and even maintains their basic credit score unchanged in extremely adverse environments, thereby avoiding the wrongful rejection of high-quality customers and demonstrating the rationality and adaptability of the algorithm.

[0127] The method of this invention determines the initial correction factor by mapping the total dynamic matching score to the matching level range. When the total dynamic matching score is lower than the preset standard and the environmental impedance index exceeds the preset impedance threshold, the deviation is determined to be a rational risk-avoidance behavior and a rational compensation value is generated for correction. This achieves dynamic adjustment of the basic credit score and avoids the wrong rejection of high-quality customers due to environmental factors.

[0128] In one embodiment, the environmental impedance index is calculated based on real-time three-stream data, and the environmental impedance index is positively correlated with the scarcity of goods, the degree of road congestion, and the degree of energy replenishment congestion.

[0129] Among them, the scarcity of goods is the ratio of the baseline order density to the real-time order density; the degree of road congestion is the ratio of the designed traffic speed to the real-time average vehicle speed; and the degree of refueling congestion is the ratio of the real-time queuing time to the baseline tolerable time.

[0130] The environmental impedance index is a weighted sum of the scarcity of goods, the degree of road congestion, and the degree of energy replenishment congestion.

[0131] In practical applications, the calculation logic of the environmental impedance index is as follows: First, define the scarcity of goods, which is inversely proportional to the order density in real-time freight flow data; second, define the road congestion, which is inversely proportional to the traffic efficiency in real-time vehicle flow data; third, define the energy replenishment congestion, which is directly proportional to the queuing time in real-time energy flow data.

[0132] The specific calculation logic is as follows:

[0133] 1. Scarcity of supply : ,in This serves as the benchmark order density for the same period in the region's history. This represents the current order density in real-time freight flow data.

[0134] 2. Road congestion : ,in The standard speed for road design. This reflects the current traffic efficiency (actual average vehicle speed) as shown by real-time traffic flow data.

[0135] 3. Energy replenishment congestion level : ,in This represents the current average queuing time at charging stations. Set a baseline queuing time that is tolerable for users.

[0136] 4. Environmental impedance index The comprehensive calculation formula is as follows:

[0137]

[0138] in, This is the environmental weighting coefficient, and .

[0139] The method of this invention identifies rational risk-averse behavior by setting a preset impedance threshold, avoiding misjudgments caused by environmental factors such as severe weather, traffic congestion, or energy price fluctuations, and providing technical support for the accurate identification of high-quality customers.

[0140] Step 103: Based on the dynamic score of the declaration credibility, execute the first cash flow simulation and the second cash flow simulation in parallel to generate the first simulated cash flow with the operation declaration information as input and the second simulated cash flow with the counterfactual parameters corrected by real-time three-stream data as input.

[0141] In one embodiment, the first cash flow simulation and the second cash flow simulation are performed in parallel, including:

[0142] Using the fare, operating frequency, and route energy consumption in the operation declaration information as input parameters, combined with the vehicle fixed costs, the net cash flow during the financing cycle is extrapolated according to the time series to generate the first simulated cash flow.

[0143] Distortion parameters in the operational declaration information that differ from real-time three-stream data by more than a preset threshold are identified. The corresponding parameters in the real-time three-stream data are used to replace the distortion parameters to obtain counterfactual parameters. Using the counterfactual parameters as input, combined with the vehicle fixed costs, the net cash flow within the financing cycle is extrapolated according to the same time series to generate a second simulated cash flow.

[0144] In practical applications, the specific execution process of dual-track simulation is as follows:

[0145] 1. First-track simulation (ideal declaration scenario): The server uses the user-submitted operational declaration information as the core input parameter, combined with regional static benchmark data (such as vehicle theoretical energy consumption and average maintenance costs), to construct the first financial model. Specifically, the first financial model is a discounted cash flow (DCF) variant structure—a net cash flow forecasting model. Its structure includes:

[0146] Revenue forecasting module (freight rates) mileage The system includes a frequency and expenditure calculation module (containing fixed costs such as insurance, dynamic costs such as electricity and tolls), as well as a time series output matrix divided by month.

[0147] In this model, the system assumes that users operate strictly according to the declared routes, frequencies, and fares. The system extrapolates monthly, calculating monthly revenue (fare × mileage), expenses (electricity + tolls + maintenance + insurance), and net cash flow for each month within a future financing cycle (e.g., 36 months). This sequence of data constitutes the first simulated cash flow. This represents the user's expected optimistic scenario.

[0148] 2. Second-Track Simulation (Counterfactual / Real-World Scenario): The server initiates error correction mode. First, it identifies parameters in the operational declaration information that differ from real-time three-flow data by more than a preset threshold and marks them as distorted parameters. For example, if a user declares a freight rate of 10 yuan / km, but real-time freight flow data shows that the average freight rate in the area is only 6 yuan / km and has been declining over the past year, then the freight rate is identified as a distorted item. Then, the server replaces those distorted parameters with real-time three-flow data (such as the actual freight rate of 6 yuan / km and the actual low turnover rate caused by traffic congestion), constructing a counterfactual scenario. Under this scenario, the system again extrapolates monthly, and the calculated net cash flow sequence constitutes the second simulated cash flow. This represents the actual cash flow level that the user might achieve under objective operational constraints, and is usually conservative.

[0149] The method of this invention distinguishes between the application scenario and the real-world scenario through parallel simulation and comparative calculation, enabling the risk assessment server to perform stress tests on the future repayment ability of the financing applicant, thus providing a basis for subsequent judgment of the risk level of credit granting.

[0150] Step 104: Based on the degree of deviation between the first simulated cash flow and the second simulated cash flow, and combined with the dynamic score of the declaration credibility, generate a differentiated credit granting strategy that includes credit limit and interest rate.

[0151] In one embodiment, based on the degree of deviation between the first simulated cash flow and the second simulated cash flow, and combined with the dynamic score of the declaration credibility, a differentiated credit granting strategy including credit limit and interest rate is generated, including:

[0152] Calculate the sum of the absolute values ​​of the differences between the first simulated cash flow and the second simulated cash flow at each time point, and calculate the proportion of the sum of the absolute values ​​of the differences to the cumulative total of the second simulated cash flow to obtain the core deviation index.

[0153] If the core deviation indicator exceeds the preset deviation tolerance threshold, it is determined that there is a significant distortion in the operational declaration information, and the second simulated cash flow is used as the credit assessment benchmark; otherwise, the integration weight of the first simulated cash flow and the second simulated cash flow is determined according to the core deviation indicator, and the weighted integration result is used as the credit assessment benchmark.

[0154] The credit limit and interest rate are determined based on the dynamic score of the application's credibility and the credit assessment benchmark.

[0155] In practical use, the server needs to quantify the difference between the two simulated cash flows to determine whether the user is excessively embellishing the application materials.

[0156] 1. Calculating the Core Deviation Index: The server first calculates the absolute value of the difference between the first and second simulated cash flows at each simulation node (i.e., each month). Then, it sums up the absolute values ​​of the differences for all months to obtain the total absolute deviation. Next, it divides the total absolute deviation by the total amount (or the total absolute value) of the second simulated cash flow (i.e., the real-world scenario) to obtain a percentage value. This value is the core deviation index. Formulaic description: The core deviation index equals (the sum of the absolute values ​​of the differences between the first and second simulated cash flows) divided by (the sum of the second simulated cash flows).

[0157] 2. Forced acceptance logic: The server will calculate the core deviation indicators. Deviation tolerance threshold (For example, 20%) will be compared, and the server will ultimately base its decision on the core deviation metrics. The first and second simulated cash flow series are unified and transformed into the final accepted cash flow series. Specifically:

[0158] Assume the planned financing period is Month, the Months ( The first simulated cash flow is The second simulated cash flow is The final accepted cash flow is .

[0159] (1) If the core deviation index Less than the threshold This indicates that although the user's report may have some discrepancies, it is generally reliable. The system may comprehensively consider both sets of data, and in this case, the final accepted cash flow will be... The calculation logic is as follows:

[0160]

[0161] in For the first simulated cash flow (user declaration), For the second simulated cash flow (real-world scenario), The trust weight coefficient is calculated using the following formula: Core deviation index. The smaller the deviation, the higher the weight given to the user's reported data.

[0162] (2) It should also be noted that if the core deviation index Exceeding the preset deviation tolerance threshold The system determines that the operational declaration information contains significant distortions. At this point, the system triggers a mandatory acceptance mechanism, completely discarding the user's declared first simulated cash flow and mandating that subsequent credit granting strategies be based solely on the second simulated cash flow (the real-world scenario). This effectively prevents users from fraudulently obtaining high credit lines by falsely reporting income. At this point:

[0163] .

[0164] The method of this invention calculates the core deviation index between the first simulated cash flow and the second simulated cash flow, and forces the adoption of the second simulated cash flow when the core deviation index exceeds the preset deviation tolerance threshold; otherwise, it performs weighted fusion based on the core deviation index to achieve differentiated determination of credit assessment benchmarks, providing a basis for the generation of credit limits and interest rates.

[0165] In one embodiment, the credit limit and interest rate are determined based on the dynamic credibility score of the declaration and the credit assessment benchmark, including:

[0166] The cash flow coverage ratio and cash flow volatility index are calculated based on the credit assessment benchmark. The cash flow coverage ratio is the ratio of the cumulative value of the credit assessment benchmark to the cumulative value of the outstanding principal and interest, and the cash flow volatility index is the ratio of the standard deviation to the mean of the credit assessment benchmark.

[0167] A three-dimensional comprehensive score is calculated by combining the dynamic score of the credibility of the application, the cash flow coverage ratio, and the cash flow volatility index.

[0168] The three-dimensional comprehensive score is mapped to a preset risk classification table to determine the main risk level and volatility sub-level.

[0169] The basic credit limit range is determined based on the main risk level, and the interest rate fluctuation ratio is determined based on the volatility sub-level, thus generating a differentiated credit strategy.

[0170] In practical use, the steps to generate a differentiated credit granting strategy are as follows:

[0171] 1. Calculate the three-dimensional comprehensive score: The server weights and integrates the dynamic score of the declaration credibility obtained in the previous steps (representing the person), the cash flow coverage rate obtained from the simulation (representing money), and the cash flow fluctuation index (representing stability) to calculate a three-dimensional comprehensive score of 0-100.

[0172] The calculation steps for cash flow coverage ratio and cash flow volatility index are as follows:

[0173] Unified in the fifth stage After finally accepting the cash flow series, the server calculates the cash flow coverage ratio and cash flow volatility index used to generate the credit granting strategy:

[0174] Let the monthly principal and interest repayment amount under the credit line the user intends to apply for be... ;

[0175] ① Cash flow coverage ratio: This indicator is used to quantitatively assess whether the total net income that a user can actually obtain throughout the entire financing cycle is sufficient to repay the loan principal and interest.

[0176] Specifically, it involves summing up the final accepted cash flows for each month within the forecast period and then dividing by the total principal and interest amount to be repaid to financial institutions within the same period.

[0177] A ratio greater than 1 indicates basic repayment ability; the higher the ratio, the thicker the asset safety cushion. Specific calculation formula:

[0178]

[0179] ② Cash Flow Volatility Index: Assessing only the overall coverage rate within the assessment period is insufficient to mitigate the risk of payment disruptions due to sudden drops in revenue in individual months. Therefore, this scheme introduces the coefficient of variation from statistics, i.e., the ratio of the standard deviation to the mean, as the cash flow volatility index to measure the dispersion and stability of the final accepted cash flows for each month.

[0180] This index eliminates the interference of absolute amount size and purely reflects the degree of fluctuation in returns. The smaller the index, the more stable the net income in each month; the larger the index, the more volatile the fluctuation, and the system will allocate a lower volatility sub-level and require a higher interest rate premium accordingly.

[0181] Specific calculation formula:

[0182] First, calculate the monthly average of the final accepted cash flows over the predetermined financing period. :

[0183]

[0184] Secondly, calculate the standard deviation of the sequence. :

[0185]

[0186] Finally, calculate the cash flow volatility index. :

[0187]

[0188] The above weighted fusion weight range: dynamic weighting of credibility. Coverage weight Volatility Index Weights The sum is 1.

[0189] For example: User credibility score is 85, cash flow coverage mapping score is 90, and cash flow volatility index mapping score is 75. The system assigns weights. Then the three-dimensional comprehensive score point.

[0190] 2. Mapping Risk Levels: Map the three-dimensional comprehensive score to a preset risk classification table (e.g., Level A, Level B, Level C, Level D).

[0191] The risk grading table is based on a three-dimensional comprehensive score (out of 100):

[0192] Category A (85-100 points): Extremely high-quality customers. Corresponding to high loan amounts (highest loan-to-value ratio) and interest rates below the benchmark rate.

[0193] Category B (70-84 points): Good customer. Corresponds to standard credit limit and benchmark interest rate.

[0194] Category C (60-69 points): High-risk borderline customers. Corresponding measures include credit limit reduction, such as lending only 60% of the benchmark interest rate, a significant increase in the interest rate, and the need for additional credit enhancement measures, such as a guarantor.

[0195] Category D (below 60 points): Loan application rejected outright.

[0196] 3. Determine Strategy Parameters: Based on the determined primary risk level, lock in the basic credit limit range (this range has been verified through the aforementioned circuit breaker mechanism). Determine the base interest rate fluctuation ratio based on the volatility sub-level (i.e., the stability of cash flow). Higher volatility results in a higher interest rate to cover the risk premium. The final portfolio output includes a differentiated credit strategy with specific credit limits (e.g., 750,000 yuan), annualized interest rates (e.g., 4.5%), and repayment periods (e.g., 36 installments).

[0197] The aforementioned "cash flow volatility index" is the sole indicator used to calculate the "volatility sub-level." A higher index, indicating a larger coefficient of variation and more fluctuating monthly income, corresponds to a lower volatility sub-level. The system directly adjusts the risk premium based on this sub-level; a higher index results in a higher spread on the base interest rate fluctuation ratio.

[0198] The method of this invention calculates the cash flow coverage ratio and cash flow volatility index, and calculates a three-dimensional comprehensive score by combining the dynamic score of the declaration credibility, the cash flow coverage ratio and the cash flow volatility index. Then, the three-dimensional comprehensive score is mapped to a risk classification table to determine the main risk level and the volatility sub-level, thereby realizing the differentiated determination of the credit limit range and the interest rate fluctuation ratio.

[0199] In one embodiment, before generating the differentiated credit policy, the method further includes:

[0200] Extract environmental characteristic factors from operational declaration information and real-time three-stream data. Environmental characteristic factors include at least ambient temperature difference, road condition slope, and load.

[0201] The scenario-based depreciation acceleration factor is calculated based on environmental characteristic factors to predict the residual value of assets at the end of the financing cycle; among them, the scenario-based depreciation acceleration factor is positively correlated with the superimposed effects of environmental temperature difference, road condition slope and load.

[0202] If the remaining value of the assets is lower than the preset lower limit of the basic credit line, the credit line circuit breaker mechanism will be triggered, and the basic credit line range will be adjusted downward based on the remaining value of the assets. The adjusted credit line range will be used as a constraint for generating differentiated credit strategies.

[0203] In this embodiment, an asset value circuit breaker mechanism is introduced before generating the final credit granting strategy. This is to prevent the risk of default caused by a borrower having good credit but a significant depreciation in the value of the vehicle assets.

[0204] Specifically, the server performs the following sub-steps:

[0205] 1. Extracting Environmental Characteristic Factors: The server further analyzes operational declaration information and real-time three-stream data to extract environmental characteristic factors that specifically affect the lifespan of power batteries. These include at least:

[0206] Environmental temperature difference factor: the deviation between the annual average temperature of the target route and the battery's optimal operating temperature (e.g., 25°C), as well as extreme low temperature data in winter. Low temperature is a major factor leading to accelerated battery degradation.

[0207] Road condition gradient factor: the average elevation change rate of the target route. Frequent heavy-load uphill climbs will accelerate battery aging through high-rate discharge.

[0208] Load factor: The ratio of estimated load rate to the vehicle's rated load capacity. Prolonged full load or overload will exacerbate mechanical wear and battery stress.

[0209] 2. Calculation of Scenario-Based Depreciation Acceleration Factor: The server uses the factors mentioned above and, based on the electrochemical lifetime model, calculates the scenario-based depreciation acceleration factor. For example, the calculation logic for this factor is: base factor (usually 1) multiplied by (1 + temperature difference influence weight × temperature difference value) multiplied by (1 + slope influence weight × slope value) multiplied by (1 + load influence weight × load value). The weights are determined through multiple linear regression analysis of historical power battery degradation data. The temperature difference influence weight typically ranges from [value missing]. The weighting range of slope influence is within The load impact weight range is within .

[0210] If the vehicle is running a courier service in a plain and temperate area, the coefficient may be close to 1; if the vehicle is hauling coal in a frigid mountainous area, the coefficient may be as high as 2.5.

[0211] Example Supplement: Assume a vehicle is hauling coal in a frigid mountainous area. Ambient temperature difference (deviation from optimal temperature) )for The average gradient of the road is Overload rate (i.e., a ratio of 1.2). The weights of each factor are taken at their extreme values. .

[0212] Then the scenario-based depreciation acceleration factor (Approximately 2.5).

[0213] 3. Predicting Residual Asset Value and Circuit Breaker: The server uses a scenario-based depreciation acceleration factor to predict the exponential decay of vehicle value, calculating the residual asset value at the end of the finance lease period (e.g., the 36th month). Then, the server compares this predicted residual value with the lower limit of the basic credit line range in the proposed credit plan (or the estimated outstanding principal balance).

[0214] Exponential decay prediction logic:

[0215] Using the improved exponentially decreasing model, the formula can be:

[0216]

[0217] Set the value of the new car The industry benchmark monthly depreciation constant is 800,000 yuan. Financing cycle Months, calculated as follows:

[0218]

[0219] The system predicted that the vehicle would only be worth about 212,800 yuan in three years, and based on this, it accurately triggered the limit circuit breaker mechanism.

[0220] Circuit breaker trigger: If the remaining value of an asset falls below the lower limit, i.e., a value inversion occurs, the system will immediately trigger the limit circuit breaker mechanism.

[0221] Credit Limit Adjustment: The system adjusts the basic credit limit range downward based on the predicted residual value of the assets. For example, if the original plan was to borrow 800,000, but it is predicted that the car will only be worth 200,000 after 3 years (while the principal will still be 300,000), the system will forcibly reduce the loan limit to 700,000 or require an increase in the down payment ratio, thereby eliminating risk exposure.

[0222] The method of this invention extracts environmental characteristic factors and calculates scenario-based depreciation acceleration coefficients to predict the residual value of assets at the end of the financing cycle. When the residual value of assets is lower than the preset lower limit of the basic credit line, a credit line circuit breaker mechanism is triggered to adjust the range of the basic credit line downward. This realizes the verification of asset value constraints and provides an upper limit constraint for the generation of differentiated credit strategies.

[0223] In one embodiment, the above method further includes:

[0224] Cryptographic hashing is performed on the operational declaration information, the data summary of real-time three-stream data, the first simulated cash flow, the second simulated cash flow, and the differentiated credit granting strategy to generate a unique digital fingerprint;

[0225] Digital fingerprints and timestamps are written into the blockchain evidence storage system to solidify the data of the entire pre-loan assessment process.

[0226] Specifically, the operational declaration information (original vouchers), the data summary of real-time three-stream data (verification basis), the calculation results of the dual-track simulation process (process data), and the finally generated differentiated credit granting strategy (result data) are packaged into a data package.

[0227] The server then uses a hash algorithm (such as SHA-256) to generate a unique digital fingerprint (hash value) for the data packet.

[0228] Finally, the digital fingerprint and the timestamp of its generation are written into the blockchain storage system of the financial consortium blockchain. This allows for verification of the authenticity and integrity of the data upon which the credit decision is based during subsequent review or evidence presentation, preventing data tampering.

[0229] The method of this invention generates a unique digital fingerprint by performing cryptographic hash operations on the data digests of operational declaration information, real-time three-stream data, the first simulated cash flow, the second simulated cash flow, and the differentiated credit granting strategy. The digital fingerprint and timestamp are then written into a blockchain storage system, realizing the solidification and storage of data throughout the entire pre-loan assessment process. This ensures the immutability and traceability of the assessment data throughout the entire process.

[0230] This invention provides a pre-loan risk assessment method for new energy commercial vehicles. First, it acquires the operational declaration information of the financing applicant and collects real-time three-flow data that has a spatiotemporal correspondence with the operational declaration information. The real-time three-flow data includes real-time freight flow data, real-time vehicle flow data, and real-time energy flow data. Then, it calculates the total dynamic matching score between the operational declaration information and the real-time three-flow data. Based on the total dynamic matching score and the environmental resistance index indicated by the real-time three-flow data, it corrects the financing applicant's basic credit score to obtain a dynamic score for declaration credibility. The environmental resistance index is used to quantify the severity of the current operating environment. Next, based on the dynamic score for declaration credibility, it executes a first cash flow simulation and a second cash flow simulation in parallel, generating a first simulated cash flow with operational declaration information as input and a second simulated cash flow with counterfactual parameters corrected by the real-time three-flow data as input. Finally, based on the deviation between the first and second simulated cash flows and combined with the dynamic score for declaration credibility, it generates a differentiated credit granting strategy that includes credit limits and interest rates. This invention's method, by introducing real-time freight flow, vehicle flow, and energy flow data, achieves a shift from "static observation of people to dynamic observation of scenarios." It solves the credit risk problem in existing new energy commercial vehicle financing and leasing businesses caused by the disconnect between static declaration information and dynamic operating environment, and the mismatch between asset value and operating scenario. Under the premise of ensuring credit security, it can accurately identify rational risk-averse behaviors and significantly improve the robustness of the risk control model and the accuracy of credit decisions.

[0231] Based on the above Figure 1 The corresponding embodiment describes a pre-loan risk assessment method for new energy commercial vehicles. Another embodiment of the present invention also provides a pre-loan risk assessment device for new energy commercial vehicles. This device includes a processor and a memory. The memory stores at least one computer instruction, which is loaded and executed by the processor to achieve the above-described method. Figure 1 The corresponding embodiment describes the pre-loan risk assessment method for new energy commercial vehicles.

[0232] Based on the above Figure 1The corresponding embodiment of the new energy commercial vehicle pre-loan risk assessment method described in the embodiments of the present invention also provides a computer-readable storage medium. For example, a non-transitory computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a CD-ROM, magnetic tape, a floppy disk, and an optical data storage system, etc. The storage medium stores at least one computer instruction for executing the above-described method. Figure 1 The pre-loan risk assessment method for new energy commercial vehicles described in the corresponding embodiments will not be repeated here.

[0233] Other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the disclosure herein. This application is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of the invention are indicated by the claims.

[0234] It should be understood that the present invention is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A method for pre-loan risk assessment of new energy commercial vehicles, characterized in that, The method includes: Obtain operational declaration information from financing applicants and collect real-time three-flow data that have a spatiotemporal correspondence with the operational declaration information; the real-time three-flow data includes real-time freight flow data, real-time vehicle flow data, and real-time energy flow data; The dynamic matching score between the operational declaration information and the real-time three-stream data is calculated, and the basic credit score of the financing applicant is corrected based on the dynamic matching score and the environmental resistance index indicated by the real-time three-stream data to obtain the dynamic credibility score of the declaration; wherein, the environmental resistance index is used to quantify the severity of the current operating environment, and is calculated based on the real-time three-stream data, and is positively correlated with the scarcity of goods, the degree of road congestion, and the degree of energy replenishment congestion; Based on the dynamic score of the declaration credibility, the first cash flow simulation and the second cash flow simulation are executed in parallel to generate the first simulated cash flow with the operational declaration information as input and the second simulated cash flow with the counterfactual parameters after being corrected by the real-time three-stream data as input. Based on the degree of deviation between the first simulated cash flow and the second simulated cash flow, and combined with the dynamic score of the declaration credibility, a differentiated credit granting strategy including credit limit and interest rate is generated. The step of adjusting the applicant's basic credit score based on the dynamic matching score and the environmental resistance index indicated by the real-time three-stream data to obtain a dynamic credit score for the application includes: The dynamic matching score is mapped to a preset matching level range to determine the corresponding initial correction factor; wherein, when the dynamic matching score is lower than the preset standard, the initial correction factor is negative. When the total dynamic matching score is lower than the preset standard and the environmental impedance index exceeds the preset impedance threshold, it is determined that the deviation between the operational declaration information and the real-time three-stream data belongs to rational risk avoidance behavior. A rational compensation value is generated based on the environmental impedance index, and the rational compensation value is used to correct the initial correction factor to obtain the final correction factor, and the final correction factor is not lower than zero; otherwise, the initial correction factor is used as the final correction factor. The final correction factor is added to the basic credit score to obtain the dynamic score of the declaration credibility.

2. The pre-loan risk assessment method for new energy commercial vehicles according to claim 1, characterized in that, The real-time three-stream data collected, which has a spatiotemporal correspondence with the operational reporting information, includes: Determine the planned operation route and work area in the operation application information; Using the planned operation route and work area as the spatial reference and the current time or the reporting plan cycle as the time reference, the real-time order density and freight rate within the planned operation route and work area are retrieved from the data interface as real-time freight flow data, the operation density and traffic efficiency of the same type of vehicles are retrieved as real-time vehicle flow data, and the idle rate and energy price of charging and swapping facilities are retrieved as real-time energy flow data.

3. The pre-loan risk assessment method for new energy commercial vehicles according to claim 1, characterized in that, The calculation of the total dynamic matching score between the operational declaration information and the real-time three-stream data includes: The matching scores for each item in the freight flow dimension are determined based on the deviation between the declared freight rate and the real-time freight rate, as well as the deviation between the declared frequency and the real-time order density; wherein, the deviation is the ratio of the absolute difference between the declared value and the real-time value to the real-time value. The traffic flow dimension matching score is determined based on the spatial overlap ratio between the declared route and the mainstream trajectory of similar vehicles; wherein, the spatial overlap ratio is the ratio of the length of the overlapping road segment to the total length of the declared route. The sub-matching score of the energy flow dimension is determined based on the deviation between the declared energy replenishment time and the actual waiting time estimated based on the real-time status of charging and swapping facilities. Obtain the preset weight coefficients for each dimension, and sum the weighted scores of each sub-item matching score to obtain the total dynamic matching score.

4. The pre-loan risk assessment method for new energy commercial vehicles according to claim 1, characterized in that, The scarcity of goods is the ratio of the baseline order density to the real-time order density; the road congestion is the ratio of the designed traffic speed to the real-time average vehicle speed; and the refueling congestion is the ratio of the real-time queuing time to the baseline tolerable time. The environmental impedance index is a weighted sum of the scarcity of goods, the degree of road congestion, and the degree of energy replenishment congestion.

5. The pre-loan risk assessment method for new energy commercial vehicles according to claim 1, characterized in that, The parallel execution of the first cash flow simulation and the second cash flow simulation includes: Using the fare, operating frequency, and route energy consumption in the operational declaration information as input parameters, and combined with the vehicle fixed costs, the net cash flow during the financing cycle is extrapolated according to the time series to generate the first simulated cash flow. Distortion parameters in the operational declaration information that differ from the real-time three-stream data by more than a preset threshold are identified. The corresponding parameters in the real-time three-stream data are used to replace the distortion parameters to obtain counterfactual parameters. Using the counterfactual parameters as input, combined with the vehicle fixed costs, the net cash flow within the financing cycle is extrapolated according to the same time series to generate the second simulated cash flow.

6. The pre-loan risk assessment method for new energy commercial vehicles according to claim 1, characterized in that, The step involves generating a differentiated credit granting strategy, including credit limit and interest rate, based on the deviation between the first simulated cash flow and the second simulated cash flow, and in conjunction with the dynamic score of the declared credibility. Calculate the sum of the absolute values ​​of the differences between the first simulated cash flow and the second simulated cash flow at each time point, and calculate the proportion of the sum of the absolute values ​​of the differences to the cumulative total of the second simulated cash flow to obtain the core deviation index; If the core deviation index exceeds the preset deviation tolerance threshold, it is determined that the operational reporting information has significant distortion, and the second simulated cash flow is used as the credit assessment benchmark; otherwise, the fusion weight of the first simulated cash flow and the second simulated cash flow is determined according to the core deviation index, and the weighted fusion result is used as the credit assessment benchmark. The credit limit and interest rate are determined based on the dynamic score of the declared credibility and the credit assessment benchmark.

7. The pre-loan risk assessment method for new energy commercial vehicles according to claim 6, characterized in that, The process of determining the credit limit and interest rate based on the dynamic score of the declared credibility and the credit assessment benchmark includes: The cash flow coverage ratio and cash flow volatility index are calculated based on the credit assessment benchmark. The cash flow coverage ratio is the ratio of the cumulative value of the credit assessment benchmark to the cumulative value of principal and interest owed. The cash flow volatility index is the ratio of the standard deviation to the mean of the credit assessment benchmark. A three-dimensional comprehensive score is calculated by combining the dynamic score of the declaration credibility, the cash flow coverage ratio, and the cash flow fluctuation index. The three-dimensional comprehensive score is mapped to a preset risk classification table to determine the main risk level and the volatility sub-level. The basic credit limit range is determined based on the primary risk level, and the interest rate fluctuation ratio is determined based on the secondary volatility level. These factors are then combined to generate the differentiated credit strategy.

8. The pre-loan risk assessment method for new energy commercial vehicles according to claim 7, characterized in that, Before generating the differentiated credit granting strategy, the following is also included: Extract environmental characteristic factors from the operational declaration information and the real-time three-stream data. The environmental characteristic factors include at least ambient temperature difference, road condition slope, and load. The scenario-based depreciation acceleration factor is calculated based on the environmental characteristic factors to predict the residual value of assets at the end of the financing cycle; wherein, the scenario-based depreciation acceleration factor is positively correlated with the superimposed effect of the environmental temperature difference, the road slope, and the load. If the remaining value of the asset is lower than the preset lower limit of the basic credit line, the credit line circuit breaker mechanism is triggered, and the basic credit line range is adjusted downward based on the remaining value of the asset. The adjusted credit line range is used as a constraint to generate the differentiated credit strategy.

9. The pre-loan risk assessment method for new energy commercial vehicles according to claim 1, characterized in that, The method further includes: The operation declaration information, the data digest of the real-time three-stream data, the first simulated cash flow, the second simulated cash flow, and the differentiated credit granting strategy are subjected to cryptographic hashing to generate a unique digital fingerprint; The digital fingerprint and timestamp are written into the blockchain evidence storage system to solidify the data of the entire pre-loan assessment process.

Images