A vehicle insurance risk identification method, device, equipment and storage medium

Abstract

This application provides a method, apparatus, device, and storage medium for identifying vehicle insurance risks, relating to the field of information technology. The method includes: acquiring objective vehicle status data and corresponding maintenance data. The objective vehicle status data includes at least one faulty component and its corresponding fault status information, while the maintenance data includes at least one component to be repaired and its corresponding maintenance behavior information. Each component to be repaired in the maintenance data is matched with the faulty component in the objective vehicle status data to obtain a matching result. Based on the matching result, a quantitative matching degree is determined between the maintenance data and the objective vehicle status data. Based on the quantitative matching degree, risk indication information is generated to support risk assessment.

Description

Technical Field

[0001] This application relates to the field of information technology, and in particular to a method, apparatus, device and storage medium for identifying vehicle insurance risks. Background Technology

[0002] Malicious exaggeration of losses in auto insurance fraud is a long-standing core pain point in the industry. The resulting additional compensation costs account for 10-15% of the total compensation amount, seriously affecting the profits of insurance companies.

[0003] Currently, image recognition technology is typically used to compare vehicle exterior damage with repair records to initially identify fraudulent repairs. The core logic of this method is to acquire images of the accident vehicle's exterior, extract the location and area of ​​damage, and match them with the corresponding repair records for exterior parts in the repair log.

[0004] However, this technology is limited to comparing visible damage on the exterior of the vehicle, making it completely unable to detect malicious acts of escalation. Summary of the Invention

[0005] This application provides a vehicle insurance risk identification method, apparatus, equipment, and storage medium for in-depth verification of real fault data inside the vehicle and maintenance behavior to identify malicious damage expansion behavior.

[0006] Firstly, this application provides a method for identifying vehicle insurance risks. This method is applied to an electronic device. The subject executing this method can be an electronic device, a component or device applied to the electronic device (e.g., a processor, chip, or chip system), or a logic module or software capable of implementing all or part of the functions of the electronic device, including: Acquire the vehicle's objective status data and corresponding maintenance data; wherein, the vehicle's objective status data includes at least one faulty component and its corresponding fault status information, and the maintenance data includes at least one component to be repaired and its corresponding maintenance behavior information; Each component to be repaired in the maintenance data is matched with the faulty component in the vehicle's objective status data to obtain the matching result; Based on the matching results, determine the quantitative matching degree between the maintenance data and the vehicle's objective condition data; Based on the quantitative matching degree, risk indication information is generated to support risk assessment.

[0007] In the first aspect, by acquiring objective vehicle status data and corresponding maintenance data—where the objective vehicle status data includes at least one faulty component and its corresponding fault status information, and the maintenance data includes at least one component to be repaired and its corresponding maintenance behavior information—dual-source information collection of the vehicle's actual condition and maintenance requests is achieved. Matching each component to be repaired in the maintenance data with the faulty component in the vehicle's objective status data establishes a correlation between maintenance behavior and the vehicle's actual fault at the component level. Based on this matching result, a quantitative matching degree between the maintenance data and the vehicle's objective status data is determined, and risk indication information is generated to support risk assessment, ensuring that the risk assessment process is based on objective and quantitative data comparison. Since each component to be repaired needs to be matched and verified against its corresponding actual fault, any maintenance items exceeding the scope of the actual fault will be reflected in the matching results, thus affecting the quantitative matching degree and the final risk indication information. Therefore, it can effectively identify risky behaviors such as replacing undamaged parts and false repairs without faults, providing a scientific and objective basis for insurance underwriting or claims review.

[0008] In conjunction with the first aspect, in one possible implementation, risk indication information for supporting risk assessment is generated based on the quantified matching degree, including: Based on the quantitative matching degree, the amplification coefficient is calculated to characterize the maintenance behavior information corresponding to the maintenance behavior that exceeds the scope of necessary fault maintenance. Risk indication information is generated based on the amplification coefficient and the quantification matching degree.

[0009] In this implementation, based on the determined quantitative matching degree, a damage amplification coefficient is further calculated to characterize the maintenance behavior information corresponding to the maintenance behavior exceeding the necessary fault repair scope. Risk indication information is then generated based on the damage amplification coefficient and the quantitative matching degree. By introducing the damage amplification coefficient as a quantitative indicator, the degree to which maintenance behavior exceeds the reasonable fault repair scope can be digitally expressed, allowing risk identification to penetrate to the quantitative level. Since the damage amplification coefficient and the quantitative matching degree jointly participate in the generation of risk indication information, when there are clearly damage amplification items in the maintenance behavior that exceed the necessary scope, even if some maintenance items match the fault, the abnormal increase in the damage amplification coefficient will be reflected in the risk indication information, thereby accurately identifying malicious damage amplification behaviors such as exaggerating the fault severity and overhauling minor faults.

[0010] In conjunction with the first aspect, in one possible implementation, the amplification coefficient is calculated based on the quantized matching degree, including: Based on the quantitative matching degree, and combined with the proportion of the repair costs associated with the unmatched repair parts in the total repair cost, the amplification coefficient is calculated.

[0011] In this implementation, the calculation of the amplification coefficient is based on the quantitative matching degree and the proportion of repair costs associated with unmatched repaired parts in the total repair cost. Since unmatched repaired parts directly correspond to items that appear on the repair list but are not faulty in the actual vehicle condition, incorporating the proportion of repair costs associated with these suspicious items into the amplification coefficient calculation allows the coefficient to directly reflect the cost weight of unreasonable repair items in the overall repair plan. When the cost proportion of unreasonable repair items is high, the amplification coefficient increases accordingly, and vice versa. This achieves accurate quantification of the degree of amplification, directly linking risk identification results to compensation costs and providing a decision-making basis for insurance risk management.

[0012] In conjunction with the first aspect, in one possible implementation, risk indication information is generated based on the amplification coefficient and the quantification matching degree, including: The quantization matching degree is compared with the matching degree threshold, and the expansion loss coefficient is compared with the expansion loss threshold. Based on the comparison results of the quantitative matching degree and the matching degree threshold, as well as the comparison results of the damage expansion coefficient and the damage expansion threshold, the risk level corresponding to the maintenance data is determined, and risk indication information containing the risk level is generated.

[0013] In this implementation, the risk level of maintenance data is determined by comparing the quantitative matching degree with a matching degree threshold and the amplification coefficient with the amplification threshold, based on the combination of the two comparison results. Risk indication information containing the risk level is then generated. Because both quantitative matching degree and amplification coefficient are used for threshold comparison simultaneously, the risk level determination process forms a two-dimensional cross-validation mechanism: a quantitative matching degree below the threshold indicates a problem in the correspondence between the maintenance item and the fault, while an amplification coefficient above the threshold indicates that the maintenance behavior has exceeded the necessary scope. When both dimensions point to high risk, the determination result has higher credibility; even when only one dimension is abnormal, a corresponding suspected risk warning can be generated. This risk level classification method based on dual threshold comparison avoids the one-sidedness of single-indicator judgment, enabling accurate classification of risk cases of different severity levels, and providing a clear and objective classification basis for subsequent differentiated handling measures.

[0014] In conjunction with the first aspect, in one possible implementation, the matching result includes first matching dimension data and second matching dimension data, wherein: The first matching dimension data is used to indicate whether each component to be repaired in the maintenance data matches a corresponding faulty component in the vehicle's objective status data; The second matching dimension data is used to indicate the compatibility information between the maintenance behavior recorded in the maintenance behavior information and the fault status indicated by the fault status information in each pair of parts that are successfully matched in the first matching dimension data. The quantitative matching degree is calculated based on the first matching dimension data and the second matching dimension data.

[0015] In this implementation, the matching results are split into first-dimensional matching data and second-dimensional matching data. The first-dimensional matching data indicates whether each component to be repaired in the maintenance data matches a corresponding faulty component in the vehicle's objective state data. The second-dimensional matching data indicates the compatibility between the maintenance behavior recorded in the maintenance behavior information and the fault state indicated in the fault state information for each successfully matched pair of components. A quantitative matching degree is calculated based on these two dimensions. By introducing a two-dimensional matching result analysis, the calculation process of the quantitative matching degree achieves a two-layer verification of the maintenance behavior: the first dimension ensures that the maintenance project is indeed targeting a faulty component, preventing the risky behavior of replacing fault-free components; the second matching dimension ensures that the maintenance behavior taken for the faulty component matches the severity of the fault, preventing the expansion of damage through over-repair of minor faults. This layered verification mechanism allows the quantitative matching degree to comprehensively reflect the true rationality of the maintenance behavior, providing a more refined and accurate quantitative basis for risk identification.

[0016] In conjunction with the first aspect, one possible implementation involves matching the parts to be repaired in the maintenance data with the faulty parts in the vehicle's objective state data, including: Based on a pre-defined standardized parts rule library, the identification information of the parts to be repaired is mapped and compared with the identification information of the faulty parts.

[0017] In this implementation, during component matching, a pre-defined standardized component rule base is used to map and compare the identification information of the component to be repaired with that of the faulty component. Since different brands and models of vehicles have different naming conventions for their components, the same component may be given different names in different scenarios. Through the unified mapping of the standardized component rule base, various synonymous and near-synonymous component names can be associated with standardized component entities, eliminating matching barriers caused by naming differences. This rule-based mapping comparison method enables component information from different data sources to achieve accurate correspondence under a unified standard, ensuring the accuracy and consistency of the matching results. It avoids mismatches or missed matches caused by inconsistent terminology, laying a reliable data foundation for subsequent quantitative calculations of the matching degree.

[0018] In conjunction with the first aspect, in one possible implementation, the second matching dimension data is determined by comparing maintenance behavior information with corresponding standard maintenance behavior information, and the standard maintenance behavior information is recorded in the mapping rule base between faults and maintenance behaviors.

[0019] In this implementation, the second matching dimension data is determined by comparing maintenance behavior information with standard maintenance behavior information recorded in a fault-to-maintenance behavior mapping rule base. Since the mapping rule base predefines the standard maintenance behavior ranges corresponding to different fault states, the fit is high when the actual maintenance behavior falls within this range, and low when it exceeds it. By introducing this rule-based comparison mechanism, the fit judgment between maintenance behavior and fault state no longer relies on subjective experience but is based on standardized rules. This standardized comparison method ensures the objectivity and consistency of the fit judgment, enabling any excessive maintenance behavior that does not conform to conventional maintenance logic to be identified, providing accurate second matching dimension data input for the quantitative calculation of the matching degree.

[0020] In conjunction with the first aspect, in one possible implementation, the method further includes: Acquire historical data, which includes historical vehicle objective condition data, historical maintenance data, and corresponding historical risk assessment results; Establish a standardized parts rule base based on historical data.

[0021] In this implementation, historical data, including objective status data of historical vehicles, historical maintenance data, and corresponding historical risk assessment results, is acquired, and a standardized parts rule base is established based on this historical data. Since the rule base is built upon the accumulation of real historical data and feedback from historical risk assessment results, the mapping relationships within the rule base can continuously absorb and reflect the real-world situations in actual business scenarios. When new vehicle models, new parts naming methods, or new maintenance modes emerge, the rule base can be continuously optimized and improved with the continuous accumulation of historical data, resulting in a sustained increase in the accuracy and coverage of parts mapping.

[0022] Secondly, this application provides a vehicle insurance risk identification device, comprising: The data acquisition module is used to acquire the vehicle's objective status data and corresponding maintenance data; wherein, the vehicle's objective status data includes at least one faulty component and its corresponding fault status information, and the maintenance data includes at least one component to be repaired and its corresponding maintenance behavior information; The parts matching module is used to match each part to be repaired in the maintenance data with the faulty parts in the vehicle's objective status data to obtain the matching result; The data quantification module is used to determine the degree of quantitative matching between maintenance data and vehicle objective status data based on the matching results. The information generation module is used to generate risk indication information to support risk assessment based on the quantitative matching degree.

[0023] In conjunction with the second aspect, in one possible implementation, the information generation module is also used to calculate, based on the quantified matching degree, the amplification coefficient that characterizes the maintenance behavior information corresponding to the maintenance behavior exceeding the necessary fault maintenance scope. Risk indication information is generated based on the amplification coefficient and the quantification matching degree.

[0024] In conjunction with the second aspect, in one possible implementation, the information generation module is also used to calculate the damage amplification coefficient based on the quantitative matching degree and the proportion of the repair costs associated with the unmatched repaired parts in the repair data in the total repair cost.

[0025] In conjunction with the second aspect, in one possible implementation, the information generation module is also used to compare the quantized matching degree with the matching degree threshold and to compare the expansion loss coefficient with the expansion loss threshold. Based on the comparison results of the quantitative matching degree and the matching degree threshold, as well as the comparison results of the damage expansion coefficient and the damage expansion threshold, the risk level corresponding to the maintenance data is determined, and risk indication information containing the risk level is generated.

[0026] In conjunction with the second aspect, in one possible implementation, the matching result includes first matching dimension data and second matching dimension data, wherein: The first matching dimension data is used to indicate whether each component to be repaired in the maintenance data matches a corresponding faulty component in the vehicle's objective status data; The second matching dimension data is used to indicate the compatibility information between the maintenance behavior recorded in the maintenance behavior information and the fault status indicated by the fault status information in each pair of parts that are successfully matched in the first matching dimension data. The quantitative matching degree is calculated based on the first matching dimension data and the second matching dimension data.

[0027] In conjunction with the second aspect, in one possible implementation, the parts matching module is also used to map and compare the identification information of the parts to be repaired with the identification information of the faulty parts based on a preset standardized parts rule library.

[0028] In conjunction with the second aspect, in one possible implementation, the second matching dimension data is determined by comparing maintenance behavior information with corresponding standard maintenance behavior information, and the standard maintenance behavior information is recorded in the mapping rule base between faults and maintenance behaviors.

[0029] In conjunction with the second aspect, in one possible implementation, the vehicle insurance risk identification device also includes a rule building module for acquiring historical data, which includes historical vehicle objective status data, historical maintenance data, and corresponding historical risk assessment results. Establish a standardized parts rule base based on historical data.

[0030] Thirdly, this application provides an electronic device comprising: a processor and a memory; the memory storing processor-executable instructions; when the processor is configured to execute the instructions, causing the electronic device to implement the method of the first aspect described above.

[0031] Fourthly, this application provides a computer-readable storage medium comprising: computer software instructions; which, when executed in an electronic device, cause the electronic device to implement the method described in the first aspect.

[0032] Fifthly, this application provides a computer program product comprising a computer program; when the computer program is run in an electronic device, it causes the electronic device to implement the method described in the first aspect.

[0033] The beneficial effects of the second to fifth aspects mentioned above are described in the corresponding description of the first aspect and will not be repeated here. Attached Figure Description

[0034] Figure 1 This is a schematic diagram illustrating the application environment of a vehicle insurance risk identification method provided in an embodiment of this application. Figure 2 A schematic diagram of a vehicle insurance risk identification system architecture provided in this application embodiment; Figure 3 A flowchart illustrating a vehicle insurance risk identification method provided in this application embodiment; Figure 4 This is a schematic diagram of a three-dimensional mapping library structure provided in an embodiment of this application; Figure 5 A schematic diagram illustrating the composition of a vehicle insurance risk identification device provided in this application embodiment; Figure 6 This is a schematic diagram of the composition of an electronic device provided in an embodiment of this application. Detailed Implementation

[0035] The following is a detailed description, with reference to the accompanying drawings, of a vehicle insurance risk identification method, apparatus, device, and storage medium provided in this application.

[0036] In this article, the term "and / or" is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.

[0037] The terms "first" and "second," etc., used in the specification and drawings of this application are used to distinguish different objects or to distinguish different treatments of the same object, rather than to describe a specific order of objects.

[0038] Furthermore, the terms "comprising" and "having," and any variations thereof, used in the description of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the steps or units listed, but may optionally include other steps or units not listed, or may optionally include other steps or units inherent to such process, method, product, or apparatus.

[0039] It should be noted that in the embodiments of this application, the words "exemplary" or "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of the words "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0040] To facilitate a clear description of the technical solutions of the embodiments of this application, the terms "first" and "second" are used in the embodiments of this application to distinguish the same or similar items with essentially the same function and effect. Those skilled in the art can understand that the terms "first" and "second" are not intended to limit the quantity or execution order.

[0041] In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0042] As the auto insurance industry continues to expand, auto insurance fraud, especially malicious damage exaggeration, has become a core pain point affecting the healthy development of the industry. Statistics show that additional compensation costs due to malicious damage exaggeration account for as much as 10% to 15% of the total compensation payout in the auto insurance industry, placing a heavy financial burden on insurance companies. Malicious damage exaggeration typically manifests as exaggerating the extent of faults, replacing undamaged parts, or performing false repairs during vehicle repairs to fraudulently obtain high compensation payments. Such fraudulent activities not only erode the profits of insurance companies but also indirectly increase the insurance costs for all car owners.

[0043] Currently, there are two main approaches to combating auto insurance fraud. One is a macro-level screening method based on historical claims data and user credit. This method analyzes macro-level data such as the frequency of historical claims, consistency of insurance information, and blacklist records to build a rule base to identify abnormal characteristics such as "high-frequency claims." However, this method relies solely on indirect behavioral characteristics, has a single data source, and cannot reach the core correlation between the actual fault state of the vehicle and repair behavior. Therefore, it struggles to identify specific malicious damage-expanding behaviors such as "replacing undamaged parts" or "exaggerating the severity of the fault," especially lacking the ability to identify first-time fraudulent acts. Furthermore, it heavily relies on manual verification, resulting in low efficiency. The other approach is a single-image recognition comparison scheme for external damage. This method collects images of vehicle exterior damage and performs a preliminary comparison with records of exterior parts in the repair list. However, this method has an extremely limited scope, only covering visible external damage such as bumpers and headlights. It is completely ineffective against fraudulent activities involving internal components such as engines, transmissions, and, in new energy vehicles, core components like battery packs, motors, and electronic control systems. Meanwhile, image recognition results are easily affected by environmental factors such as shooting angle and lighting, and there is a lack of standardized rules that associate "damage level" with "repair actions", resulting in high false positive and false negative rates.

[0044] This demonstrates that existing technologies generally suffer from drawbacks such as limited data sources, lack of standardized comparison logic, difficulty in quantifying the extent of damage expansion, and insufficient adaptability to new scenarios like new energy vehicles. The root cause lies in the industry's absence of a mechanism for in-depth field-level verification of objective vehicle condition data and repair behavior data. The lack of a unified mapping relationship and quantitative standards between "parts-fault status-repair behavior" makes it impossible to accurately and automatically determine the rationality of repair lists, hindering the formation of effective technical defenses against malicious damage expansion.

[0045] To address the aforementioned technical problems, this application provides a method, apparatus, device, and storage medium for identifying vehicle insurance risks. The approach involves acquiring objective vehicle status data and corresponding maintenance data. The objective status data includes at least one faulty component and its corresponding fault status information, while the maintenance data includes at least one component requiring repair and its corresponding maintenance behavior information. This achieves dual-source information collection of the vehicle's actual condition and maintenance requests. Matching each component requiring repair in the maintenance data with the faulty component in the vehicle's objective status data establishes a correlation between maintenance behavior and the vehicle's actual fault at the component level. Based on this matching result, a quantitative matching degree is determined between the maintenance data and the vehicle's objective status data. Risk indication information is then generated based on this quantitative matching degree to support risk assessment, ensuring that the risk assessment process is grounded in objective and quantitative data comparison. Since each component requiring repair needs to be matched and verified against a corresponding actual fault, any maintenance items exceeding the scope of the actual fault will be reflected in the matching results, thus affecting the quantitative matching degree and the final risk indication information. Therefore, this approach effectively identifies risky behaviors such as replacing undamaged parts and false repairs without faults, providing a scientific and objective basis for insurance underwriting or claims review.

[0046] The embodiments provided in this application will now be described in detail with reference to the accompanying drawings.

[0047] The vehicle insurance risk identification method provided in this application can be applied to, for example... Figure 1 The application environment shown. For example... Figure 1 As shown, the application environment includes: Terminal device 100 and server 101.

[0048] The terminal device 100 includes an application 102 that supports vehicle insurance risk identification. The client of the application 102, which supports vehicle insurance risk identification, is used in the terminal device 100 to visualize risk indication information during the execution of the vehicle insurance risk identification method according to this embodiment.

[0049] This client provides a user interface, which can take the form of a World Wide Web (Web) page accessed through a browser or a native application that needs to be downloaded and installed. The terminal is specifically a user equipment (UE), which includes, but is not limited to, smartphones, tablets, laptops, desktop computers, and Internet of Things (IoT) terminals. The terminal accesses the access network via a wireless air interface, possessing the capability to carry voice services, data transmission services, and multimedia services. It can also achieve direct communication between different terminals based on device-to-device (D2D) direct connection technology.

[0050] This client is used to receive data query requests from users and display risk indication information to users.

[0051] In another example, the vehicle insurance risk identification method provided in this application can be applied to server 101. Server 101 runs an application 102 that supports vehicle insurance risk identification. This application is responsible for processing requests sent by clients and executing the vehicle insurance risk identification method, which includes: obtaining the vehicle's objective state data and corresponding maintenance data; matching the parts to be repaired in the maintenance data with the faulty parts in the vehicle's objective state data to obtain a matching result; determining the quantitative matching degree between the maintenance data and the vehicle's objective state data based on the matching result; and generating risk indication information to support risk assessment based on the quantitative matching degree.

[0052] In one alternative embodiment, the terminal device 100 and the server 101 can be interconnected via a wired or wireless network.

[0053] Server 101 includes a first memory and a first processor. The first memory stores a vehicle insurance risk identification program; the vehicle insurance risk identification program is invoked and executed by the first processor to implement the vehicle insurance risk identification method provided in this application. The first memory may include, but is not limited to, the following: random access memory (RAM), read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), and electrically erasable programmable read-only memory (EEPROM). The first processor may consist of one or more integrated circuit chips. Optionally, the first processor may be a general-purpose processor, such as a central processing unit (CPU) or a network processor (NP). Optionally, the first processor can implement the vehicle insurance risk identification method provided in this application by running programs or code.

[0054] This application embodiment also provides a vehicle insurance risk identification system, which can be set up in... Figure 1 In the application environment shown, such as Figure 2 As shown, the vehicle insurance risk identification system 200 may include: Input layer 201: Used to acquire the vehicle's objective status data and corresponding maintenance data; wherein, the vehicle's objective status data includes at least one faulty component and its corresponding fault status information, and the maintenance data includes at least one component to be repaired and its corresponding maintenance behavior information. Processing layer 202: Used for in-depth analysis and risk assessment of the acquired data. Processing layer 202 further includes: The Parts Matching Engine 2021 is used to match each part to be repaired in the maintenance data with the faulty parts in the vehicle's objective state data to obtain a matching result. The matching result includes first matching dimension data and second matching dimension data. The first matching dimension data indicates whether each part to be repaired in the maintenance data matches a corresponding faulty part in the vehicle's objective state data (e.g., mapping and comparing based on part model, installation location, and other identification information). The second matching dimension data indicates the compatibility information between the maintenance behavior recorded in the maintenance behavior information and the fault state indicated by the fault state information in each pair of parts successfully matched in the first matching dimension data. The Matching Engine 2021 includes a first matching dimension data calculation unit and a second matching dimension data calculation unit. The first matching dimension data calculation unit maps and compares the identification information of the parts to be repaired with the identification information of the faulty parts based on a standardized parts rule base to determine whether the parts are successfully matched. The second matching dimension data calculation unit compares the maintenance behavior information in successfully matched parts pairs with standard maintenance behavior information based on a fault and maintenance behavior mapping rule base to determine the compatibility information between the maintenance behavior and the fault state.

[0055] Data Quantization Engine 2022: Used to calculate the quantitative matching degree between maintenance data and vehicle objective status data based on first and second matching dimension data. Data Quantization Engine 2022 includes a quantitative matching degree calculation unit, which fuses the component matching results corresponding to the first matching dimension data with the adaptability information corresponding to the second matching dimension data, outputting a quantitative matching degree with a value ranging from 0 to 1.

[0056] Coefficient Calculation Engine 2023: The Coefficient Calculation Engine 2023 includes a loss amplification coefficient calculation unit. This unit calculates the loss amplification coefficient based on the quantified matching degree. The loss amplification coefficient characterizes the degree to which the repair behavior corresponding to the repair behavior information exceeds the necessary fault repair scope. Based on the quantified matching degree, and combined with the proportion of the repair cost associated with the unmatched repaired parts in the repair data in the total repair cost, the loss amplification coefficient is calculated.

[0057] Risk Assessment Engine 2024: Used to generate risk indication information based on the damage amplification coefficient and the quantified matching degree. Specifically, it compares the quantified matching degree with a preset matching degree threshold and the damage amplification coefficient with a preset damage amplification threshold; based on the comparison results, it determines the risk level (e.g., normal, suspected fraud, confirmed fraud) corresponding to the maintenance data and generates risk indication information containing the risk level.

[0058] Rule Building Engine 2025: Used to acquire historical data and build or update a standardized component rule base and a fault and maintenance behavior mapping rule base based on the historical data.

[0059] Output layer 203: Used to output risk indication information. Output layer 203 may include a result display interface, data interface, and alarm generation unit, facilitating auditors to quickly obtain the basis for judgment and take corresponding measures.

[0060] It should be noted that the system architecture described in the embodiments of this application is for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and does not constitute a limitation on the technical solutions provided in the embodiments of this application. As those skilled in the art will know, with the evolution of system architecture, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

[0061] See Figure 3 This is a flowchart illustrating a vehicle insurance risk identification method provided in an embodiment of this application. Figure 3 As shown, the vehicle insurance risk identification method provided in this application can be implemented through the aforementioned server, specifically including the following steps S300~S303.

[0062] S300: The server obtains the vehicle's objective status data and corresponding maintenance data.

[0063] The vehicle's objective status data includes at least one faulty component and its corresponding fault status information, while the maintenance data includes at least one component to be repaired and its corresponding maintenance action information. The fault status information describes the specific fault manifestations, extent of damage, or performance parameters of the faulty component. The maintenance action information describes the maintenance operations to be performed on the component to be repaired.

[0064] Vehicle objective status data, for example, refers to vehicle operating status data directly acquired through vehicle-to-everything (V2X) data acquisition devices. These devices can be vehicle-mounted telematics boxes (T-Boxes), battery management systems (BMS) for new energy vehicles, or onboard sensors (e.g., collision sensors, component status sensors, performance monitoring sensors, etc.). After a vehicle accident, the acquisition devices automatically collect data in real time, with a collection frequency set to once every 5 minutes after the accident, continuously for 24 hours. This ensures both the timeliness of the data (avoiding distortion due to changes in fault status over time) and the completeness of the data (covering the dynamic changes in component status after the accident). The collected content includes component information (e.g., component model, factory serial number, installation location, etc. associated with the vehicle identification number (VIN) as component identifiers), fault status (e.g., fault codes, fault levels (e.g., minor, moderate, severe), performance degradation rate, self-test pass rate, etc.), accident-related sensor data, and accident-related sensor data such as collision intensity, collision location, and real-time operating parameters of components after the accident.

[0065] Repair data, such as unprocessed vehicle repair records obtained from the business systems of repair service providers, is synchronized in real time with these systems via a standardized application programming interface (API). It supports automatic parsing of repair lists in mainstream formats such as portable document format (PDF) and Excel spreadsheets, eliminating the need for manual data entry. The collected data covers three dimensions: repair object, repair behavior, and parts attributes. Repair object data includes information such as the model, specifications, or installation location of parts to be replaced or repaired. Repair behavior data includes specific actions such as disassembly, replacement, inspection, and cleaning. Part attributes data includes part type (e.g., original equipment manufacturer (OEM), aftermarket, used parts), part specifications, and a description of the reason for replacement. The repair data also includes detailed repair time records to help determine the reasonableness of the repair behavior.

[0066] These two types of data together constitute a complete description of the actual fault status of the vehicle and the planned maintenance behavior, providing an objective basis for risk assessment.

[0067] S301. The server matches each component to be repaired in the maintenance data with the faulty component in the vehicle's objective status data to obtain the matching result.

[0068] The server matches each component requiring repair recorded in the maintenance data with the faulty components recorded in the vehicle's objective state data to determine the correspondence between the repair request and the actual fault. The matching process uses the component's unique identifier information, such as model number, code, or installation location, to compare each component in the maintenance data with the actual faulty component in the vehicle's objective state data. This matching operation yields the matching result for each component requiring repair. Establishing this correspondence provides a foundation for subsequently quantifying the degree of matching between maintenance actions and fault states, thus supporting an objective judgment of the rationality of the repair.

[0069] S302. Based on the matching results, the server determines the quantitative matching degree between the maintenance data and the vehicle's objective status data.

[0070] Based on the matching results, the server determines a quantitative matching degree to characterize the overall consistency between maintenance data and vehicle objective condition data. The quantitative matching degree is a numerical indicator ranging from 0 to 1, reflecting the level of agreement between the maintenance requests in the maintenance list and the actual fault condition of the vehicle. A value closer to 1 indicates a higher degree of matching between the maintenance action and the fault state; a value closer to 0 indicates a greater discrepancy. The matching results include the correspondence between each component to be repaired and the faulty component in the maintenance data. Based on these correspondences, the server integrates multi-dimensional comparison information into a comprehensive quantitative value using a pre-set statistical algorithm or model, thus transforming complex comparison conclusions into a comparable and verifiable numerical form. This quantitative matching degree provides an objective quantitative basis for subsequent assessment of maintenance rationality and identification of potential risks, elevating the judgment of maintenance actions from qualitative analysis to quantitative calculation.

[0071] S303. The server generates risk indication information to support risk assessment based on the quantified matching degree.

[0072] The server generates risk indication information based on quantified matching degree to support risk assessment. This risk indication information is a comprehensive evaluation result of the matching relationship between maintenance data and vehicle objective condition data. It is used to alert the risk assessment system or reviewers to potential anomalies in maintenance activities, assisting subsequent risk identification decisions. The quantified matching degree, expressed as a value between 0 and 1, reflects the degree of agreement between the maintenance request and the actual fault. Based on the range of this value or a comparison with a preset reference value, the server outputs corresponding risk indication information according to a predetermined risk assessment logic. This information may include risk warning markers, abnormal item indications, or suggested attention levels, transforming abstract numerical values ​​into intuitive decision-making references. This provides objective evidence for claims review, risk investigation, or manual review, thereby achieving intelligent assistance in insurance risk management.

[0073] In this embodiment, by acquiring vehicle objective status data and corresponding maintenance data, where the vehicle objective status data includes at least one faulty component and its corresponding fault status information, and the maintenance data includes at least one component to be repaired and its corresponding maintenance behavior information, dual-source information collection of the vehicle's actual condition and maintenance requests is achieved. Matching each component to be repaired in the maintenance data with the faulty component in the vehicle objective status data establishes a correlation between maintenance behavior and the vehicle's actual fault at the component level. Based on this matching result, a quantitative matching degree between the maintenance data and the vehicle objective status data is determined, and risk indication information is generated to support risk assessment, ensuring that the risk assessment process is based on objective and quantitative data comparison. Since each component to be repaired needs to be matched and verified against its corresponding actual fault, any maintenance items exceeding the scope of the actual fault will be reflected in the matching results, thus affecting the quantitative matching degree and the final risk indication information. Therefore, it can effectively identify risky behaviors such as replacing undamaged parts and false repairs without faults, providing a scientific and objective decision-making basis for insurance underwriting or claims review.

[0074] In one embodiment, step S303 includes: S3031. Based on the quantified matching degree, the server calculates the amplification coefficient, which is used to characterize the maintenance behavior information corresponding to the maintenance behavior, if the maintenance behavior exceeds the scope of necessary fault maintenance.

[0075] The damage amplification coefficient is a numerical indicator used to quantify the degree of deviation between unreasonable repairs and reasonable repairs in the repair list. Its value is typically non-negative; a higher value indicates a more severe degree of damage amplification. The server obtains the quantitative matching degree determined through the aforementioned matching process, reflecting the degree of agreement between the repair request and the actual fault. A lower quantitative matching degree indicates a lower degree of agreement between the fault and the repair, and a greater likelihood of damage amplification. Based on the quantitative matching degree, and combined with other parameters related to the reasonableness of the repair behavior, the server uses a pre-defined damage amplification coefficient calculation model to transform the degree of damage amplification from a qualitative description into a quantifiable value, providing a clear quantitative basis for subsequent risk assessment.

[0076] S3032. The server generates risk indication information based on the amplification coefficient and the quantification matching degree.

[0077] The server takes the calculated damage amplification coefficient and quantitative matching degree as input, and generates corresponding risk indication information based on preset risk assessment logic. This information may include risk level identifiers, abnormal item markings, or suggested handling methods, used to alert the risk assessment system or reviewers to potential anomalies in the repair process. By combining the damage amplification coefficient and quantitative matching degree to generate risk indication information, the rationality and potential risks of the repair process can be more comprehensively reflected, providing objective support for subsequent claims decisions.

[0078] In one possible implementation, step S3031 includes: S30311. The server calculates the damage amplification coefficient based on the quantitative matching degree and the proportion of the repair costs associated with the unmatched repair parts in the total repair cost.

[0079] Unmatched repair components refer to repair list items that, during the component matching process, failed to correspond to any faulty components in the vehicle's objective condition data. These items may be considered fault-free repairs or fictitious repairs. Unmatched repair items directly reflect unnecessary repair activities, and the proportion of their associated repair costs in the total repair cost is an important quantitative indicator for measuring the degree of damage amplification. Based on the quantified matching degree and the proportion of repair costs associated with unmatched repair components in the repair data within the total repair cost, the server derives a specific value using a pre-defined damage amplification coefficient calculation model. For example, the damage amplification coefficient can be calculated using the formula... L =(1- S )×(1+ D )× K Calculation, where S This represents the quantitative matching degree, with a value ranging from 0 to 1; D The depreciation rate of representative parts is calculated based on the service life of the vehicle. For example, the depreciation rate of a vehicle used for 3 years is 30%. The higher the depreciation rate, the more obvious the damage attribute of unnecessary replacement. K The component replacement rationality coefficient reflects the compatibility between the parts used in the repair and the original vehicle, as well as the necessity of replacement. For example, original parts are preferred when replacement is necessary. K =1, when replacing aftermarket or used parts is not necessary. K The value ranges from 1.2 to 1.5, with larger values ​​indicating higher levels of risk amplification. The amplification coefficient calculated using this formula transforms the risk amplification behavior from a qualitative description into a quantitative value, providing a clear quantitative basis for subsequent risk classification.

[0080] In one possible implementation, a machine learning model is used to directly quantify the degree of risk amplification. The server uses quantitative matching degree, component depreciation rate, accessory type, repair hours, accessory price, and fault level as input features, and trains an XGBoost, Random Forest, or Logistic Regression model based on historical risk case data and normal case data. The trained model directly outputs the probability of risk amplification, ranging from 0 to 1. This model can automatically learn the non-linear relationships between features, achieving higher quantification accuracy than linear formulas. The output probability of risk amplification is functionally equivalent to the risk amplification coefficient, and both can serve as a quantitative basis for risk assessment. For example, a probability of risk amplification greater than or equal to 0.6 corresponds to confirmed risk, between 0.3 and 0.6 corresponds to suspected risk, and less than 0.3 corresponds to normal risk.

[0081] In another possible implementation, an optimized loss amplification coefficient formula is used to calculate the degree of loss amplification. Based on the original loss amplification coefficient formula, a maintenance time rationality coefficient is introduced, and the loss amplification coefficient calculation formula is as follows: L =(1- S )×(1+ D )× K × H ,in S To quantify the degree of matching, D For component depreciation rate, K To determine the reasonableness of parts replacement, H This is the maintenance labor hour rationality coefficient. H It equals the actual repair time divided by the standard repair time. When the actual repair time significantly exceeds the standard repair time (e.g., ... H A value greater than or equal to 1.5 indicates that the working hours are artificially inflated, which exacerbates the degree of loss.

[0082] In one possible implementation, step S3032 includes: S30321. The server compares the quantization matching degree with the matching degree threshold, and compares the expansion loss coefficient with the expansion loss threshold.

[0083] The quantitative matching degree is compared with a matching degree threshold, and the damage amplification coefficient is compared with the damage amplification threshold. The matching degree threshold is a pre-set critical value used to judge the consistency between maintenance data and vehicle objective state data. For example, a first matching degree threshold of 0.5 and a second matching degree threshold of 0.8 are set to identify low and medium matching degrees, respectively. The damage amplification threshold is a pre-set critical value used to judge the severity of damage amplification. For example, a first damage amplification threshold of 0.3 and a second damage amplification threshold of 0.6 are set to identify no damage amplification, suspected damage amplification, and obvious damage amplification, respectively. The server compares the calculated quantitative matching degree S with 0.5 and 0.8 respectively to determine the interval in which S falls; simultaneously, it compares the damage amplification coefficient L with 0.3 and 0.6 respectively to determine the interval in which L falls. Through this dual comparison, the comparison results of two dimensions are obtained, providing a basis for subsequent comprehensive risk level determination. This comparison process transforms continuous numerical values ​​into discrete interval classifications, facilitating the establishment of clear risk determination rules.

[0084] S30322. Based on the comparison results of the quantitative matching degree and the matching degree threshold, and the comparison results of the damage expansion coefficient and the damage expansion threshold, the server determines the risk level corresponding to the maintenance data and generates risk indication information containing the risk level.

[0085] Based on the comparison results of the quantitative matching degree and the matching degree threshold, and the comparison results of the amplification coefficient and the amplification threshold, the risk level corresponding to the maintenance data is determined, and risk indication information containing the risk level is generated. The risk level is a classification identifier for the potential risks of maintenance data, such as dividing it into three levels: normal maintenance, suspected risk, and confirmed risk. The server classifies the risk level based on the preset dual-threshold cross-judgment rules: when the quantitative matching degree S is greater than or equal to 0.8 and the amplification coefficient L is less than 0.3, it is judged as normal maintenance; when S is greater than or equal to 0.5 and less than 0.8 and L is greater than or equal to 0.3 and less than 0.6, or S is less than 0.5 but L is less than 0.3, or S is greater than 0.8 but L is greater than 0.6, etc., the judgment logic can be refined according to the actual business rules. For example, the case where S is between 0.5 and 0.8 or L is between 0.3 and 0.6 is judged as suspected risk; when S is less than 0.5 and L is greater than or equal to 0.6, it is judged as confirmed risk. The server generates corresponding risk indication information based on the judgment results. This information clearly includes a risk level identifier, such as "normal," "suspected," or "high risk," and can be accompanied by a brief explanation. For suspected risk levels, the risk indication information can also suggest key items to focus on or dimensions to be reviewed. By generating risk indication information that includes risk levels, complex comparisons and calculations are transformed into intuitive decision-making guidance, facilitating claims reviewers to quickly take appropriate measures and achieving automated and precise risk management. Items marked as normal are directly approved for claims without manual intervention, improving claims efficiency. Items marked as suspected automatically trigger a manual review process, simultaneously pushing dual-source data, matching results, similarity, and damage amplification coefficient calculation processes to reviewers to assist in rapid manual judgment. For items marked as high risk, claims for corresponding unreasonable repair items are automatically rejected, and the vehicle owner and repair service provider are blacklisted, restricting their future insurance and claims privileges.

[0086] One possible implementation involves using dynamic thresholds instead of fixed thresholds for risk grading. The server uses a sliding window technique to statistically analyze the risk rate distribution across different regions and vehicle types quarterly, dynamically adjusting the quantitative matching degree threshold and the amplification coefficient threshold based on the statistical results. For example, for regions with high risk rates, the quantitative matching degree threshold is raised to 0.6, and the amplification coefficient threshold is lowered to 0.5. Alternatively, a K-means clustering algorithm can be used, employing the quantitative matching degree and amplification coefficient of historical cases as clustering features to automatically divide the data into three clusters: "normal," "suspected," and "high-risk," replacing manually set fixed thresholds. This approach adapts to different scenarios and achieves the core function of risk grading without requiring manual threshold setting.

[0087] In another possible implementation, Bayesian probability judgment is used instead of threshold rules for risk classification. The server calculates prior probabilities, such as the risk rate P(F) of a certain car model, based on historical data; and calculates conditional probabilities, such as the risk probability P(S<0.5|F) when the quantitative matching degree S is less than 0.5. Based on the Bayesian formula P(F|S,L)=P(S,L|F)×P(F) / P(S,L), the posterior risk probability is calculated, where P(F|S,L) represents the probability of a risk event occurring given the quantitative matching degree S and the amplification coefficient L. The risk is classified according to the range of the posterior risk probability; for example, a probability greater than or equal to 0.8 is considered confirmed fraud, a probability between 0.3 and 0.8 is considered suspected risk, and a probability less than 0.3 is considered normal.

[0088] In this embodiment, based on the determined quantitative matching degree, a damage amplification coefficient is further calculated to characterize the maintenance behavior information corresponding to the maintenance behavior exceeding the necessary fault repair scope. Risk indication information is then generated based on the damage amplification coefficient and the quantitative matching degree. By introducing the damage amplification coefficient as a quantitative indicator, the degree to which maintenance behavior exceeds the reasonable fault repair scope can be digitally expressed, allowing risk identification to penetrate to a quantitative level. Since the damage amplification coefficient and the quantitative matching degree jointly participate in the generation of risk indication information, when there are damage amplification items that clearly exceed the necessary scope in the maintenance behavior, even if some maintenance items match the fault, the abnormal increase in the damage amplification coefficient will be reflected in the risk indication information, thereby accurately identifying malicious damage amplification behaviors such as exaggerating the fault severity and overhauling minor faults.

[0089] When calculating the amplification coefficient, it is based on the quantitative matching degree and combined with the proportion of repair costs associated with unmatched repaired parts in the total repair cost. Since unmatched repaired parts directly correspond to items that are not faulty in the actual vehicle condition but appear on the repair list, by including the proportion of repair costs associated with these suspicious items in the calculation of the amplification coefficient, the coefficient can intuitively reflect the cost weight of unreasonable repair items in the overall repair plan. When the cost proportion of unreasonable repair items is high, the amplification coefficient increases accordingly, and vice versa. This achieves accurate quantification of the degree of amplification, directly linking risk identification results with compensation costs, and providing a decision-making basis for insurance risk management.

[0090] By comparing the quantitative matching degree with a matching degree threshold and the amplification coefficient with the amplification threshold, the risk level corresponding to the maintenance data is determined based on the combination of the two comparison results, and risk indication information containing the risk level is generated. Because both quantitative matching degree and amplification coefficient are used for threshold comparison simultaneously, the risk level determination process forms a two-dimensional cross-validation mechanism: a quantitative matching degree below the threshold indicates a problem in the correspondence between the maintenance item and the fault, while a amplification coefficient above the threshold indicates that the maintenance behavior has exceeded the necessary scope. When both dimensions point to high risk, the determination result has higher credibility; even when only one dimension is abnormal, a corresponding suspected risk warning can still be generated. This risk level classification method based on dual threshold comparison avoids the one-sidedness of single-indicator judgment, enabling accurate classification of risk cases of different severity levels, and providing a clear and objective classification basis for subsequent differentiated handling measures.

[0091] In one embodiment, the matching result includes first matching dimension data and second matching dimension data, wherein: The first matching dimension data is used to indicate whether each component to be repaired in the maintenance data matches a corresponding faulty component in the vehicle's objective state data. The server first calculates the first matching dimension data, for example using a cosine similarity algorithm. It converts the features of successfully matched components (such as component model, installation location, and fault type code) into high-dimensional feature vectors. The cosine of the angle between these vectors is used to measure the consistency between them, resulting in a first matching dimension data value ranging from 0 to 1. The closer the value is to 1, the higher the matching degree at the component level. The cosine similarity algorithm formula is as follows:

[0092] In the formula, x i For the objective state data of the vehicle, feature vectors of the components. y i This is the feature vector of the maintenance data components.

[0093] The second matching dimension data is used to indicate the compatibility information between the maintenance behavior recorded in the maintenance behavior information and the fault state indicated in the fault state information for each pair of parts that are successfully matched in the first matching dimension data. Based on the first matching dimension data, the server introduces the second matching dimension data for weighted correction to reflect the compatibility between fault severity and maintenance actions. Specifically, fault level weights are assigned according to the fault level (e.g., severe, moderate, minor) in the fault state information, for example, severe fault weight 0.6, moderate fault weight 0.3, minor fault weight 0.1; maintenance action compatibility weights are assigned according to the degree of compatibility between the maintenance behavior and standard maintenance behavior, for example, complete compatibility weight 0.8, partial compatibility weight 0.4, and incompatible compatibility weight 0. The server calculates the product of the fault level weight and the maintenance action compatibility weight, and takes the average over all successfully matched part pairs to obtain a correction coefficient.

[0094] The quantitative matching degree is calculated based on the data from the first and second matching dimensions. The first matching dimension data is multiplied by a correction factor to obtain the final quantitative matching degree, which still ranges from 0 to 1. This quantitative matching degree integrates information from two dimensions: whether the parts match and whether the repair behavior is reasonable. This allows the matching results to more comprehensively reflect the degree of agreement between the repair request and the actual fault condition, providing a precise quantitative basis for subsequent risk assessment.

[0095] In one possible implementation, the server calculates the first matching dimension data by replacing cosine similarity with Euclidean distance or Jaccard coefficient.

[0096] Euclidean distance measures the degree of matching between component features (model, installation location, fault type code) by calculating the geometric distance between feature vectors in high-dimensional space; the closer the distance, the higher the matching degree. Jaccard coefficient measures the degree of overlap at the set level by calculating the ratio of the intersection to the union of the component feature set or fault type set; the closer the coefficient is to 1, the higher the matching degree. The server uses entropy weighting or a sliding window combined with entropy weighting to replace fixed weights for correction. Entropy weighting automatically assigns weights based on the degree of variation of each indicator (such as fault level, maintenance action adaptability); the greater the degree of variation, the more information the indicator provides, and the higher the assigned weight, avoiding manual setting bias. The sliding window combined with entropy weighting dynamically captures changes in the importance of indicators over different time periods, allowing weight allocation to adapt to the evolution of insurance fraud methods.

[0097] In another possible implementation, the induced ordered weighted geometric averaging (IOWGA) operator is used to weight and combine multiple similarity algorithms to calculate the quantized matching degree.

[0098] The server calculates the first matching dimension data using cosine similarity, Euclidean distance, and Jaccard coefficient, obtaining multiple similarity calculation results. The calculation accuracy of each similarity algorithm is used as an inductive value; a higher inductive value indicates higher accuracy of the algorithm in historical data, and is therefore assigned a larger weight. These results are then combined using an inductively weighted geometric mean operator to obtain a comprehensive quantitative matching degree. This approach improves the stability of similarity calculation through multi-algorithm fusion, avoiding the problem of a single algorithm performing poorly under specific data features. This aligns with the goal of "weighted correction to improve accuracy" and is adaptable to application scenarios with different data characteristics.

[0099] In one possible implementation, step S301 includes: S3011. The server maps and compares the identification information of the parts to be repaired with the identification information of the faulty parts based on a preset standardized parts rule library.

[0100] A standardized parts rule base is a pre-built set of mapping relationships used to unify the differences in parts naming across different data sources. It records standard identification information for each part, such as standard model, standard name, or standard code, as well as the correspondence between different representations. The identification information of the parts to be repaired is extracted from the repair data, including the part model, name, or code recorded in the repair list; the identification information of faulty parts is extracted from the vehicle's objective status data, including the part's factory serial number, model, or installation location code recorded in the vehicle network data. The server matches the original identification information of the parts to be repaired with the standardized parts rule base, converts it into a standard identifier, and then compares it with the standard identifier of the faulty parts to determine if they point to the same part. Through this mapping comparison process, the correspondence between repair requests and actual faulty parts can be accurately established, eliminating data heterogeneity problems caused by inconsistent parts naming across different vehicle models. This provides an accurate parts matching basis for subsequent matching degree calculations, thereby ensuring that risky behaviors such as replacing non-faulty parts can be effectively identified.

[0101] In one possible implementation, the second matching dimension data is determined by comparing maintenance behavior information with corresponding standard maintenance behavior information, which is recorded in a fault-to-maintenance behavior mapping rule base.

[0102] The second matching dimension data is determined by comparing maintenance behavior information with corresponding standard maintenance behavior information, which comes from a fault-to-maintenance behavior mapping rule base. This rule base is pre-established and records the correspondence between different fault states and reasonable maintenance operations. For example, for specific fault types and fault levels of specific components, the rule base specifies the maintenance actions to be performed (e.g., minor faults correspond to inspection, serious faults correspond to assembly replacement) and the allowed parts types (e.g., original parts, aftermarket parts). The server extracts maintenance behavior information corresponding to each successfully matched pair of components from the maintenance data, including the proposed maintenance actions (e.g., replacement, inspection, cleaning) and parts attributes. Simultaneously, it obtains the fault status information of the component from the vehicle's objective status data, including the fault type and fault level. The server queries the mapping rule base based on the fault status information to obtain standard maintenance behavior information for that fault state. The actual maintenance behavior information is compared with the standard maintenance behavior information to assess the degree of fit, such as determining whether the maintenance action matches the fault level and whether the parts selection conforms to specifications. Through this comparison process, the second matching dimension data is obtained, which can reflect the rationality of maintenance behavior relative to the fault state, and provide a basis for the calculation of the fault and maintenance adaptability dimension for quantifying the matching degree.

[0103] In one possible implementation, step S301 further includes: S3012. The server acquires historical data, which includes historical vehicle objective status data, historical maintenance data, and corresponding historical risk assessment results.

[0104] Historical vehicle objective status data consists of actual vehicle data collected from past accidents or malfunctions, including information on faulty components and their status, such as fault codes, fault levels, and performance degradation rates. Historical maintenance data comprises maintenance records performed on corresponding vehicles by repair shops, including information on components to be repaired and actual maintenance actions, such as repair actions, parts types, and detailed work hours. Historical risk assessment results represent the final risk conclusions in corresponding historical claims cases, such as normal repair, suspected fraud, or confirmed fraud. These conclusions originate from insurance company claims review records or manual review results. Historical data is acquired in batches from insurance company claims databases, OEM vehicle networking platforms, or repair shop business systems via standardized interfaces. After cleaning and anonymization, it is used for subsequent rule base construction and model training. The purpose of acquiring historical data is to extract statistical patterns and effective experiences from a large number of real-world cases, providing a data foundation for establishing an objective and scientific rule base, ensuring that the rule base accurately reflects reasonable behavioral patterns in real-world repair scenarios.

[0105] S3013. The server establishes a standardized component rule base based on historical data.

[0106] A standardized parts rule base is a set of mapping relationships used to unify the naming differences of parts from different data sources. It records the standard identification information of each part, such as standard model, standard name, or standard code, as well as the correspondence between different manufacturers' and vehicle models' various descriptions of the same part. The server analyzes the acquired historical data, extracting all parts identification information appearing in historical vehicle objective condition data and historical maintenance data. Through statistical clustering, text similarity matching, or ontology mapping techniques, it identifies different descriptions pointing to the same actual part; for example, "headlight," "headlamp," and "vehicle headlight" are all normalized to the same standard terminology. Simultaneously, combined with historical risk assessment results, frequently occurring parts naming methods are verified and corrected to ensure the accuracy of the rule base. The established standardized parts rule base provides a unified reference benchmark for matching parts between subsequent maintenance data and vehicle objective condition data, eliminating data heterogeneity problems caused by naming differences, and providing an accurate matching basis for identifying risky behaviors such as replacing non-faulty parts.

[0107] S3014. The server uses a rule base to map faults and maintenance behaviors based on historical data.

[0108] The fault-to-maintenance behavior mapping rule base is a knowledge base that records the correspondence between different fault states and reasonable maintenance operations. It clearly defines the reasonable and unreasonable boundaries of fault types and maintenance actions for components. The server analyzes and mines historical data based on vehicle maintenance industry standards, component fault mechanisms, and risk control experience embedded in historical data. Specifically, it extracts the fault type and fault level of faulty components from historical vehicle objective state data, extracts the actual maintenance actions and parts selections from corresponding historical maintenance data, and combines this with historical risk assessment results (e.g., normal claim cases are considered a reference for reasonable mapping, and confirmed risk cases are considered typical of unreasonable mapping) to construct reasonable mapping relationships through association rules.

[0109] One possible implementation involves constructing a three-dimensional mapping library based on a standardized component rule base and a fault and maintenance behavior mapping rule base, such as... Figure 4As shown, the 3D mapping library adopts a vertical three-layer architecture: the top layer is the component dimension, divided into general components and new energy-specific components; the middle layer is the fault state dimension, including fault type and fault level codes; and the bottom layer is the maintenance behavior dimension, including maintenance behavior and component attributes. In terms of association logic, a reasonable mapping relationship points from the component dimension to the fault state dimension, representing the possible faults of a specific component, and then from the fault state dimension to the maintenance behavior dimension, representing the reasonable maintenance actions corresponding to a specific fault. For example, "headlight (component) - damaged casing (fault type) - replace casing (maintenance action)" or "battery pack (component) - cell degradation (fault type) - cell inspection (maintenance action)" constitutes a reasonable mapping. Inappropriate mapping relationships are identified as associations that do not conform to industry standards and maintenance logic. For example, "headlight (component) - no fault (fault type) - replace assembly (maintenance action)" or "motor (component) - slight abnormal noise (fault type) - replace motor assembly (maintenance action)" will be marked as high-risk behavior.

[0110] The 3D mapping library employs standardized coding, with unique component identifiers using a combination of model and installation location codes to ensure precise matching. Fault severity codes are 01 (minor), 02 (moderate), and 03 (severe), achieving standardized quantification of fault severity. Repair action classifications clearly distinguish different repair levels, such as assembly replacement, component replacement, inspection, and cleaning. For the adaptability characteristics of new energy vehicles, the 3D mapping library includes a dedicated new energy vehicle component library, listing core components such as battery packs, drive motors, and electronic control systems, and including new energy vehicle-specific fault types such as cell abnormalities and performance degradation. Stricter repair rules are set for high-value components such as battery packs. Through the clear distinction between solid and dashed arrows, the 3D mapping library provides clear rule boundaries for system judgment, reducing subjective judgment errors. The establishment of this 3D mapping library intuitively demonstrates the three-dimensional relationship between "components-faults-repairs," solving the data heterogeneity problem across different vehicles through the coding system and unique identifiers, and highlighting the special risk control logic of core components in new energy vehicles, providing a standardized judgment basis for subsequent fault and repair compatibility verification.

[0111] During the matching process, the server employs a dual-keyword plus attribute verification matching mechanism. First, by combining the first keyword, "part model, installation location," it ensures that the matched part is the same part from the same vehicle, such as "front bumper, left front side" or "battery pack, chassis center," avoiding confusion between different parts. Second, by combining the second keyword, "fault type, repair action," it verifies the compatibility between the repair action and the fault type based on the mapping relationship recorded in the three-dimensional mapping rule base. For example, does "serious fault" correspond to the "replacement" action, and does "minor fault" correspond to the "inspection" action? Finally, attribute verification is performed to additionally verify the consistency between the part's compatible vehicle model, manufacturing year, and the actual vehicle information (obtained through VIN code parsing), eliminating false data such as "repair records for parts that do not match the vehicle model." Through this three-layer verification mechanism of "precisely locating parts, verifying fault and repair compatibility, and eliminating false records," precise one-to-one correspondence at the field level is achieved, ensuring that every repair record can be traced back to the corresponding part fault, fundamentally eliminating the risk of "fault-free repair" and "repair mismatch."

[0112] In one possible implementation, a three-dimensional rule base is automatically generated through association rule mining.

[0113] The Apriori association rule mining algorithm is used to analyze massive amounts of historical normal maintenance data, including historical vehicle objective status data and corresponding historical maintenance data. The algorithm mines frequent itemsets of "parts-fault type-maintenance action," setting a support of at least 5% and a confidence of at least 80%, and automatically generates reasonable mapping rules. For anomalous itemsets that do not conform to industry standards and maintenance logic, such as "no fault - assembly replacement," they are marked as unreasonable mappings, forming a dynamically updated rule base.

[0114] In another possible implementation, a three-dimensional rule base is constructed through knowledge graph association mapping.

[0115] A vehicle maintenance knowledge graph is constructed, with "parts," "fault types," and "maintenance actions" as core nodes and "adaptation" and "correspondence" as connecting edges. Weights are assigned to these connecting edges; for example, "serious fault - replacement" has a weight of 0.9, and "minor fault - replacement" has a weight of 0.1. The graph queries are used to match the reasonableness of maintenance actions corresponding to parts and faults, replacing the traditional rule-based comparison method.

[0116] In one possible implementation, the 3D mapping library, weights, and thresholds are updated based on the updated vehicle objective status data and corresponding maintenance data.

[0117] In one possible implementation, the 3D mapping library, fault level weight, maintenance action adaptability weight, quantification matching threshold, and damage amplification coefficient threshold are updated based on the updated vehicle objective status data and corresponding maintenance data.

[0118] The server acquires new historical data at a preset cycle, which is set to 1 to 3 months. Updated data includes vehicle objective status data, maintenance data and corresponding risk assessment results, as well as manually reviewed and corrected assessment results, subsequent accidents involving blacklisted vehicles, component data, and new risk cases. Based on this new data, the server updates the 3D mapping library.

[0119] The server also adjusts the fault level weight and maintenance action adaptability weight. For example, for core components of new energy vehicles such as battery packs, drive motors, and electronic control systems, the fault level weight is increased to 0.7 to strengthen the risk control of high-value components. Simultaneously, the server optimizes the quantization matching threshold and the amplification coefficient threshold. Based on the differences in fraud rates in different regions and vehicle models, the quantization matching threshold and the amplification coefficient threshold are fine-tuned to improve the model's adaptability to different regions and vehicle models. Through the above update mechanism, the three-dimensional rule base can continuously adapt to changes in vehicle technology iteration and risk, avoiding a decrease in recognition accuracy due to technology updates or method upgrades, ensuring long-term stability and effectiveness. In one embodiment, step S300 includes: S3001. Perform data cleaning processing on the vehicle's objective status data and maintenance data, including: unifying the naming of parts and maintenance items, and removing abnormal and duplicate data.

[0120] After acquiring vehicle objective status data and maintenance data, and before matching the parts to be repaired in the maintenance data with the faulty parts in the vehicle objective status data, the server performs data cleaning on the vehicle objective status data and maintenance data. First, the server removes outliers and duplicate data from both the vehicle objective status data and maintenance data. Outliers refer to values ​​in the vehicle objective status data that deviate from the normal range, and field values ​​in the maintenance data that exceed the reasonable range. Duplicate data refers to records collected repeatedly at the same time point in the vehicle objective status data, and identical maintenance entries being entered repeatedly in the maintenance data. For outliers in the vehicle objective status data, an interquartile range algorithm is used for identification and removal; for outliers in the maintenance data, value range verification is used for identification and removal. For duplicate data in both the vehicle objective status data and maintenance data, duplicate record checks are performed separately, identifying and deleting identical records. The purpose of removing outliers and duplicate data is to ensure the accuracy and purity of the original data and to avoid abnormal or erroneous data interfering with risk assessment.

[0121] Standardizing component names involves resolving different naming conventions for the same component in vehicle objective status data and maintenance data, such as "headlight," "headlamp," and "vehicle headlight." This is achieved by constructing a unified terminology library across vehicle models using ontology mapping technology, normalizing these synonymous names into standard terms, and eliminating naming differences. Similarly, standardizing maintenance item names involves normalizing different expressions describing the same maintenance action in the original data, such as "disassembly," "removal," and "removal," into standard terms using the unified terminology library. Furthermore, unstructured qualitative fields are encoded and mapped. For example, fault levels "minor," "moderate," and "severe" are encoded as "01," "02," and "03," respectively; maintenance items "replacement," "inspection," and "cleaning" are encoded as "10," "20," and "30," respectively; and part types "original," "aftermarket," and "discarded parts" are encoded as "001," "002," and "003," respectively, achieving uniform field format. After name standardization and encoding mapping, vehicle objective status data and maintenance data are transformed into data with consistent format and unified semantics, enabling direct comparison between the two previously heterogeneous data types.

[0122] In one possible implementation, communication with the electronic control unit (ECU) is achieved via an on-board diagnostics (OBD) system interface. This interface reads standardized diagnostic trouble codes (DTCs), real-time data streams, and fault classifications, replacing the objective vehicle status data collected through a telematics processor (T-Box) or battery management system (BMS). The OBD interface serves as the vehicle's standardized data output port, while the ECU is the core control component. Together, they acquire the operating status of various vehicle systems. Standardized fault codes identify specific fault types, real-time data streams include dynamic operating parameters such as engine speed, coolant temperature, and oxygen sensor voltage, and fault classifications differentiate between temporary and permanent faults. Maintenance data is collected using blockchain technology, storing maintenance lists and records in a distributed ledger format to ensure data immutability and traceability, replacing maintenance list data synchronized via a standardized application programming interface (API).

[0123] In another possible implementation, offline inspection reports from third-party testing agencies can replace real-time collection of some vehicle objective condition data, while data exported from the repair shop's enterprise resource planning (ERP) system can replace interface synchronization of repair list data. The offline inspection reports from third-party testing agencies include fault locations, damage levels, and core component testing data. Issued after a comprehensive inspection of the vehicle using professional testing equipment, these reports objectively reflect the vehicle's actual technical condition. The repair shop's ERP system is the core system for internal management, recording information throughout the entire repair process. Data such as repair projects, parts procurement records, and time settlement sheets can be directly exported from this system. This data covers key information such as the repair object, repair activities, and parts attributes.

[0124] In one possible implementation, term normalization is achieved through word embedding techniques in natural language processing (NLP).

[0125] Using word embedding models such as word2vec or BERT, unstructured text such as part names and repair items is converted into high-dimensional semantic vectors. The cosine similarity between vectors is calculated, and terms with a quantized matching degree reaching a preset threshold (e.g., 0.95) are automatically identified as synonyms. For example, the vector similarity between "headlight" and "headlamps" exceeds the threshold, so they are classified as the same standard term.

[0126] In this embodiment, the matching result is split into a first matching dimension data and a second matching dimension data. The first matching dimension data indicates whether each component to be repaired in the maintenance data matches a corresponding faulty component in the vehicle's objective state data. The second matching dimension data indicates the compatibility between the maintenance behavior recorded in the maintenance behavior information and the fault state indicated in the fault state information for each successfully matched pair of components. A quantitative matching degree is calculated based on these two dimensions. By introducing a two-dimensional matching result analysis, the calculation process of the quantitative matching degree achieves a two-layer verification of the maintenance behavior: the first dimension ensures that the maintenance project is indeed targeting a faulty component, preventing the risky behavior of replacing fault-free components; the second matching dimension ensures that the maintenance behavior taken for the faulty component matches the severity of the fault, preventing the expansion of damage through over-repair of minor faults. This layered verification mechanism enables the quantitative matching degree to comprehensively reflect the true rationality of the maintenance behavior, providing a more refined and accurate quantitative basis for risk identification.

[0127] During component matching, a pre-defined standardized component rule base is used to map and compare the identification information of the component to be repaired with that of the faulty component. Since different brands and models of vehicles have different naming conventions for their components, the same component may be given different names in different scenarios. Through the unified mapping of the standardized component rule base, various synonymous or near-synonymous component names can be associated with standardized component entities, eliminating matching barriers caused by naming differences. This rule-based mapping comparison method enables component information from different data sources to achieve accurate correspondence under a unified standard, ensuring the accuracy and consistency of the matching results. It avoids mismatches or missed matches caused by inconsistent terminology, laying a reliable data foundation for subsequent quantitative calculations of the matching degree.

[0128] The second matching dimension data is determined by comparing maintenance behavior information with standard maintenance behavior information recorded in a fault-to-maintenance behavior mapping rule base. Since the mapping rule base predefines the standard maintenance behavior ranges corresponding to different fault states, the fit is high when the actual maintenance behavior falls within this range, and low when it exceeds it. By introducing this rule-based comparison mechanism, the fit judgment between maintenance behavior and fault state no longer relies on subjective experience but is based on standardized rules. This standardized comparison method ensures the objectivity and consistency of the fit judgment, enabling the identification of any excessive maintenance behavior that does not conform to conventional maintenance logic, and providing accurate second matching dimension data input for the quantitative calculation of the matching degree.

[0129] By acquiring historical data, including objective status data of historical vehicles, historical maintenance data, and corresponding historical risk assessment results, a standardized parts rule base is established based on this historical data. Since the rule base is built upon the accumulation of real historical data and feedback from historical risk assessment results, the mapping relationships within the rule base can continuously absorb and reflect the real-world situations in actual business scenarios. When new vehicle models, new parts naming methods, or new maintenance modes emerge, the rule base can be continuously optimized and improved with the continuous accumulation of historical data, resulting in a sustained increase in the accuracy and coverage of parts mapping.

[0130] As can be seen, the above mainly describes the solutions provided by the embodiments of this application from a methodological perspective. To achieve the above functions, the embodiments of this application provide corresponding hardware structures and / or software modules for executing each function. Those skilled in the art should readily recognize that, in conjunction with the modules and algorithm steps of the various examples described in the embodiments disclosed herein, the embodiments of this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed by hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this invention.

[0131] This application embodiment can divide the vehicle insurance risk identification device into functional modules according to the above method example. For example, each function can be divided into a separate functional module, or two or more functions can be integrated into one processing module. The integrated module can be implemented in hardware or as a software functional module. Optionally, the module division in this application embodiment is illustrative and only represents one logical functional division; other division methods may be used in actual implementation.

[0132] In some embodiments, this application also provides a vehicle insurance risk identification device. This vehicle insurance risk identification device may include one or more functional modules for implementing the vehicle insurance risk identification method of the above method embodiments.

[0133] For example, Figure 5 This is a schematic diagram illustrating the composition of a vehicle insurance risk identification device provided in an embodiment of this application. Figure 5 As shown, the vehicle insurance risk identification device 400 includes: a data acquisition module 401, a parts matching module 402, a data quantification module 403, and an information generation module 404.

[0134] The data acquisition module 401 is used to acquire the vehicle's objective status data and corresponding maintenance data. The vehicle's objective status data includes at least one faulty component and its corresponding fault status information, while the maintenance data includes at least one component to be repaired and its corresponding maintenance behavior information.

[0135] The parts matching module 402 is used to match each part to be repaired in the maintenance data with the faulty parts in the vehicle objective status data to obtain the matching result.

[0136] The data quantization module 403 is used to determine the quantification matching degree between maintenance data and vehicle objective status data based on the matching results.

[0137] The information generation module 404 is used to generate risk indication information to support risk assessment based on the quantitative matching degree.

[0138] In one embodiment, the information generation module 404 is further configured to calculate, based on the quantified matching degree, a damage amplification coefficient that characterizes the maintenance behavior corresponding to the maintenance behavior information as exceeding the necessary fault maintenance scope.

[0139] Risk indication information is generated based on the amplification coefficient and the quantification matching degree.

[0140] In one embodiment, the information generation module 404 is further configured to calculate the damage amplification coefficient based on the quantified matching degree and in combination with the proportion of the repair costs associated with the unmatched repair parts in the repair data in the total repair costs.

[0141] In one embodiment, the information generation module 404 is further configured to compare the quantized matching degree with the matching degree threshold and the expansion loss coefficient with the expansion loss threshold.

[0142] Based on the comparison results of the quantitative matching degree and the matching degree threshold, as well as the comparison results of the damage expansion coefficient and the damage expansion threshold, the risk level corresponding to the maintenance data is determined, and risk indication information containing the risk level is generated.

[0143] In one embodiment, the matching result includes first matching dimension data and second matching dimension data, wherein: The first matching dimension data is used to indicate whether each component to be repaired in the maintenance data matches a corresponding faulty component in the vehicle's objective state data.

[0144] The second matching dimension data is used to indicate the compatibility information between the maintenance behavior recorded in the maintenance behavior information and the fault status indicated by the fault status information for each pair of parts that are successfully matched in the first matching dimension data.

[0145] The quantitative matching degree is calculated based on the first matching dimension data and the second matching dimension data.

[0146] In one embodiment, the parts matching module 402 is further configured to map and compare the identification information of the parts to be repaired with the identification information of the faulty parts based on a preset standardized parts rule library.

[0147] In one embodiment, the second matching dimension data is determined by comparing maintenance behavior information with corresponding standard maintenance behavior information, and the standard maintenance behavior information is recorded in the fault-to-maintenance behavior mapping rule base.

[0148] In one embodiment, the vehicle insurance risk identification device 400 further includes a rule building module 405 for acquiring historical data, which includes historical vehicle objective status data, historical maintenance data, and corresponding historical risk assessment results.

[0149] Establish a standardized parts rule base based on historical data.

[0150] When implementing the functions of the integrated modules described above in hardware, this embodiment of the invention provides a possible schematic diagram of the electronic device involved in the above embodiments. For example... Figure 6 As shown, the electronic device 500 includes: a processor 502, a communication interface 503, and a bus 504. Optionally, the electronic device 500 may also include a memory 501.

[0151] Processor 502 may implement or execute various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. Processor 502 may be a central processing unit, a general-purpose processor, a digital signal processor, an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It may implement or execute various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. Processor 502 may also be a combination that implements computing functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.

[0152] Communication interface 503 is used to connect to other devices via a communication network. This communication network can be Ethernet, wireless access network, wireless local area network (WLAN), etc.

[0153] The memory 501 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), disk storage medium or other magnetic storage device, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but is not limited thereto.

[0154] In one possible implementation, the memory 501 can exist independently of the processor 502. The memory 501 can be connected to the processor 502 via a bus 504 and is used to store instructions or program code. When the processor 502 calls and executes the instructions or program code stored in the memory 501, it can implement the vehicle insurance risk identification method provided in this embodiment of the invention.

[0155] In another possible implementation, the memory 501 can also be integrated with the processor 502.

[0156] Bus 504 can be an extended industry standard architecture (EISA) bus, etc. Bus 504 can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 6 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0157] Through the above description of the implementation methods, those skilled in the art can clearly understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the service calling device can be divided into different functional modules to complete all or part of the functions described above.

[0158] This application also provides a computer-readable storage medium. All or part of the processes in the above method embodiments can be executed by computer instructions instructing related hardware. The program can be stored in the aforementioned computer-readable storage medium, and when executed, it can include the processes of the above method embodiments. The computer-readable storage medium can be any of the foregoing embodiments or memory. The aforementioned computer-readable storage medium can also be an external storage device of the aforementioned service invocation device, such as a plug-in hard drive, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the aforementioned service invocation device. Further, the aforementioned computer-readable storage medium can include both internal storage units of the aforementioned service invocation device and external storage devices. The aforementioned computer-readable storage medium is used to store the aforementioned computer program and other programs and data required by the aforementioned service invocation device. The aforementioned computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

[0159] This application also provides computer instructions. All or part of the processes in the above method embodiments can be executed by computer instructions to instruct related hardware (such as computers, processors, network devices, and terminals). The program can be stored in the aforementioned computer-readable storage medium.

[0160] This application also provides a computer program product that, when run on a computer, causes the above-described method embodiments to be executed.

[0161] This application also provides a chip system. The chip system may be composed of chips or may include chips and other discrete devices, without limitation. The chip system includes a processor and a transceiver. All or part of the processes in the above method embodiments can be completed by this chip system, such as the chip system being used to implement the functions performed by the network devices or terminals in the above method embodiments.

[0162] In one possible design, the chip system further includes a memory for storing program instructions and / or data. When the chip system is running, the processor executes the program instructions stored in the memory to enable the chip system to perform the functions performed by the network device or terminal in the above method embodiments.

[0163] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for identifying vehicle insurance risks, characterized in that, include: Acquire the vehicle's objective status data and corresponding maintenance data; wherein, the vehicle's objective status data includes at least one faulty component and its corresponding fault status information, and the maintenance data includes at least one component to be repaired and its corresponding maintenance behavior information; Each component to be repaired in the maintenance data is matched with the faulty component in the vehicle's objective status data to obtain a matching result; Based on the matching results, determine the quantitative matching degree between the maintenance data and the vehicle objective status data; Based on the quantified matching degree, risk indication information is generated to support risk assessment.

2. The method according to claim 1, characterized in that, The step of generating risk indication information to support risk assessment based on the quantified matching degree includes: Based on the quantified matching degree, a loss amplification coefficient is calculated to characterize the maintenance behavior corresponding to the maintenance behavior information as exceeding the necessary fault repair scope; The risk indication information is generated based on the amplification coefficient and the quantification matching degree.

3. The method according to claim 2, characterized in that, The calculation of the amplification coefficient based on the quantized matching degree includes: Based on the quantitative matching degree, and combined with the proportion of the repair costs associated with the unmatched repair parts in the repair data in the total repair cost, the amplification coefficient is calculated.

4. The method according to claim 2 or 3, characterized in that, The step of generating the risk indication information based on the amplification coefficient and the quantification matching degree includes: The quantized matching degree is compared with the matching degree threshold, and the amplification coefficient is compared with the amplification threshold; Based on the comparison results of the quantitative matching degree and the matching degree threshold, and the comparison results of the damage expansion coefficient and the damage expansion threshold, the risk level corresponding to the maintenance data is determined, and risk indication information containing the risk level is generated.

5. The method according to claim 1, characterized in that, The matching results include first matching dimension data and second matching dimension data, wherein: The first matching dimension data is used to indicate whether each component to be repaired in the maintenance data is matched with a corresponding faulty component in the vehicle objective status data; The second matching dimension data is used to indicate the compatibility information between the maintenance behavior recorded in the maintenance behavior information and the fault state indicated by the fault state information in each pair of components that are successfully matched in the first matching dimension data. The quantitative matching degree is calculated based on the first matching dimension data and the second matching dimension data.

6. The method according to claim 5, characterized in that, The process of matching the parts to be repaired in the maintenance data with the faulty parts in the vehicle's objective condition data includes: Based on a pre-defined standardized parts rule library, the identification information of the parts to be repaired is mapped and compared with the identification information of the faulty parts.

7. The method according to claim 5, characterized in that, The second matching dimension data is determined by comparing the maintenance behavior information with the corresponding standard maintenance behavior information, which is recorded in the fault and maintenance behavior mapping rule base.

8. The method according to claim 6, characterized in that, The method further includes: Acquire historical data, including historical vehicle objective status data, historical maintenance data, and corresponding historical risk assessment results; Based on the historical data, a standardized parts rule base is established.

9. A vehicle insurance risk identification device, characterized in that, include: The data acquisition module is used to acquire the vehicle's objective status data and corresponding maintenance data; wherein, the vehicle's objective status data includes at least one faulty component and its corresponding fault status information, and the maintenance data includes at least one component to be repaired and its corresponding maintenance behavior information; The parts matching module is used to match each part to be repaired in the maintenance data with the faulty parts in the vehicle objective status data to obtain a matching result; The data quantization module is used to determine the quantization matching degree between the maintenance data and the vehicle objective status data based on the matching results. The information generation module is used to generate risk indication information to support risk assessment based on the quantified matching degree.

10. An electronic device, characterized in that, The device includes a processor and a memory, the processor being coupled to the memory; the memory is used to store computer instructions, which are loaded and executed by the processor to enable the computer device to implement the vehicle insurance risk identification method as described in any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes computer-executable instructions that, when executed on a computer, cause the computer to perform the vehicle insurance risk identification method as described in any one of claims 1 to 8.

12. A computer program product, characterized in that, The computer program product includes a computer program that, when run on an electronic device, causes the electronic device to perform the vehicle insurance risk identification method as described in any one of claims 1 to 8.

Images