Risk assessment method and device based on model context protocol, equipment and medium

Abstract

The invention discloses a risk assessment method and device based on a model context protocol, computer equipment and a computer readable storage medium, and aims to reduce the repeated development workload and shorten the system docking period through a root mechanism of an MCP, a pre-constructed tool chain and a unified multi-source heterogeneous data interface. By supporting cross-system data acquisition, dynamic data aggregation is realized, data islands are eliminated, the underwriting error rate is reduced, and the underwriting accuracy of a financial insurance system is improved. And the underwriting rule is packaged through the MCP tool chain, so that personalized risk assessment is realized, and the flexibility and expandability of the financial insurance system are improved. When the method is applied to the financial insurance business field and the medical health field, the flexibility and the expandability of a financial insurance system and a medical health system can be improved, and the underwriting accuracy of insurance business is improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent decision-making technology, and in particular to a risk assessment method, device, computer equipment and computer-readable storage medium based on a model context protocol. Background Art

[0002] Current financial insurance and healthcare systems, when underwriting auto and health insurance, must connect to disparate data sources. Independent interfaces must be developed for these heterogeneous data sources (such as hospital HIS, connected vehicle onboard devices (OBDs), and biometric databases). This not only results in high development costs and a long development time, but also leads to poor flexibility and scalability due to the customization of interfaces. Furthermore, user behavior data, device sensor data, and historical claims records are scattered across multiple systems during the insurance process, lacking real-time aggregation capabilities (for example, health insurance in the healthcare sector cannot dynamically access heart rate data from wearable devices), resulting in high underwriting error rates.

[0003] Therefore, how to solve the current problems of poor flexibility, scalability and low underwriting accuracy in financial insurance and medical health systems has become an urgent issue that needs to be addressed. Summary of the Invention

[0004] The present invention provides a risk assessment method, device, computer equipment and computer-readable storage medium based on a model context protocol to solve the technical problems of poor flexibility and scalability of current financial insurance systems and medical health care, as well as low underwriting accuracy.

[0005] In a first aspect, a risk assessment method based on a model context protocol is provided, comprising:

[0006] When receiving the insurance request from the target user, the data access boundary of at least one relevant data source is defined through the roots mechanism of the Model Context Protocol (MCP) to access multiple heterogeneous data sources;

[0007] Dynamically constructing context for each modal data corresponding to the insurance request from various relevant data sources to obtain insurance-related information for the target user;

[0008] The insurance-related information is underwritten through the MCP tool chain to generate a risk assessment result corresponding to the target user.

[0009] In a second aspect, a risk assessment device based on a model context protocol is provided, comprising:

[0010] The multi-source data access module is used to define the data access boundary of at least one relevant data source through the roots mechanism of the Model Context Protocol (MCP) upon receiving the insurance application request of the target user, so as to access multiple heterogeneous data sources;

[0011] A related information construction module is used to dynamically construct context for each modal data corresponding to the insurance request from various relevant data sources to obtain insurance related information of the target user;

[0012] The insurance risk assessment module is used to underwrite the insurance-related information through the MCP tool chain and generate a risk assessment result corresponding to the target user.

[0013] In a third aspect, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of the risk assessment method based on the model context protocol are implemented.

[0014] In a fourth aspect, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the risk assessment method based on the model context protocol are implemented.

[0015] In the scheme implemented by the risk assessment method, device, computer equipment and storage medium based on the above-mentioned model context protocol, when receiving the insurance request of the target user, the data access boundary of at least one relevant data source is defined through the roots mechanism of the model context protocol MCP to access multiple heterogeneous data sources; dynamic context is constructed for each modal data corresponding to the insurance request from each relevant data source to obtain the insurance-related information of the target user; through the MCP tool chain, the insurance-related information is underwritten to generate the risk assessment result corresponding to the target user. In the present invention, the roots mechanism of MCP and the pre-built tool chain are used to unify the multi-source heterogeneous data interface, thereby reducing the workload of repeated development and shortening the system docking cycle. By supporting cross-system data acquisition, dynamic data aggregation is achieved, data silos are eliminated, the underwriting error rate is reduced, and the underwriting accuracy of the financial insurance system is improved. The underwriting rules are encapsulated by the MCP tool chain to achieve personalized risk assessment, thereby improving the flexibility and scalability of the financial insurance system. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments of the present invention. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0017] Figure 1 This is a flow chart of a risk assessment method based on a model context protocol in one embodiment of the present invention;

[0018] Figure 2 yes Figure 1 A schematic flow chart of a specific implementation of step S10;

[0019] Figure 3 yes Figure 1 A schematic flow chart of a specific implementation of step S20;

[0020] Figure 4 yes Figure 1 A schematic flow chart of a specific implementation of step S30;

[0021] Figure 5 1 is a schematic structural diagram of a risk assessment device based on a model context protocol according to an embodiment of the present invention;

[0022] Figure 6 It is a structural diagram of a computer device in one embodiment of the present invention. DETAILED DESCRIPTION

[0023] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0024] The risk assessment method based on the model context protocol involved in the embodiment of the present application is mainly applied to computer devices. The risk assessment generation device based on the model context protocol can be a PC, a portable computer, a mobile terminal, or other device with display and processing functions.

[0025] The following describes some embodiments of the present application in detail with reference to the accompanying drawings. In the absence of conflict, the following embodiments and features therein may be combined with each other.

[0026] Reference Figure 1 , Figure 1 A flowchart of a risk assessment method based on a model context protocol provided in this application.

[0027] like Figure 1 As shown, an embodiment of the present application provides a risk assessment method based on a model context protocol, and the risk assessment method based on a model context protocol includes steps S10 to S30.

[0028] In this embodiment, the risk assessment method based on the model context protocol includes the following steps:

[0029] Step S10: upon receiving the insurance application request from the target user, defining the data access boundary of at least one relevant data source through the roots mechanism of the Model Context Protocol (MCP) to access multiple heterogeneous data sources;

[0030] Currently, traditional financial business systems have the following problems in insurance underwriting:

[0031] 1. Traditional system integration and development efficiency is low: independent interfaces need to be developed for different data sources (such as hospital HIS, car networking OBD, and credit reporting platform), resulting in interface fragmentation and poor version compatibility. Single system integration takes 3-6 months, resulting in high development costs.

[0032] 2. Slow data processing and high risk assessment error rate: User data, device sensor data, and historical claims records in the insurance process are scattered across multiple systems, lacking real-time aggregation capabilities (for example, health insurance cannot dynamically obtain heart rate data from wearable devices). Reliance on static risk assessment models and data silos result in an underwriting error rate as high as 8.7%.

[0033] 3. Poor scalability and ecological compatibility: Traditional system function upgrades require code reconstruction, and the cost of adapting to new insurance types is high.

[0034] To address the aforementioned issues, this embodiment provides a system architecture based on the Model Context Protocol (incorporating the MCP protocol stack). The Model Context Protocol (MCP) integrates large language models with external data sources and tools, establishing a secure, bidirectional connection between the large model and the data source. The system architecture is divided into three main layers: the MCP client layer, the MCP server cluster, and the intelligent decision-making layer. Each layer has different components. For example, the MCP client layer includes an insurance app and a biometric recognition module; the MCP server cluster includes a medical data server, an Internet of Vehicles server, and a blockchain evidence storage server. Through this system architecture, this embodiment provides a risk assessment method based on the Model Context Protocol. Specifically, when a user initiates insurance, the client collects basic user information and device data (such as heart rate data from wearable devices and On-Board Device (OBD) data from Internet of Vehicles) and then transmits this data to the corresponding server via the MCP protocol for processing. Each server in the server cluster (such as the medical data server and the Internet of Vehicles server) processes these requests, generates risk assessment results, and returns the results to the intelligent decision-making layer, ultimately generating the premium calculation and policy.

[0035] Among them, the insurance APP is used to provide a user interaction portal, integrating the biometric recognition model and device data collection module;

[0036] The biometric module is used for liveness detection and identity verification;

[0037] The device data collector is used to connect wearable devices (such as heart rate monitoring), vehicle networking OBD and other sensors;

[0038] Medical data server, used to connect to the hospital HIS system to provide medical history query services;

[0039] Internet of Vehicles server, used to process real-time OBD data and calculate driving scores;

[0040] Blockchain evidence storage server, used to store electronic insurance policy hashes and underwriting logs (Merkle tree structure).

[0041] Dynamic model loader, used to load risk assessment models based on MCP server capability description (JSON-LD);

[0042] A multi-tool collaboration engine, used to orchestrate the tool call sequence in the tool chain (e.g., identity verification → data query → premium calculation), and drives the process through prompts templates;

[0043] Based on the above system architecture, a risk assessment method based on model context protocol is provided, which includes a dynamic context construction engine, tool chain automated underwriting and federated learning enhanced risk assessment.

[0044] Specifically, when a user initiates an insurance request, user identity verification is first performed, such as through liveness detection for biometric identification. Specifically, user-provided biometric data (such as facial recognition or fingerprint data) is transmitted to the MCP client for verification. The MCP client can submit an authorization authentication request to the MCP server cluster using the liveness detection video stream / fingerprint feature code + user ID. After authentication, the MCP server cluster can provide the MCP client with an authentication token (JWT) and a list of authorized roots (such as medical / Internet of Vehicles data domains). This triggers the server's " / get_medical_history" tool, which then dynamically obtains user-related data (such as medical records, driving behavior, etc.) based on the user's basic information (such as age and vehicle model) used to authorize device data access (OBD, GPS). For example, by calling the " / calculate_driving_score" interface on the Internet of Vehicles server, the data collection scope can be dynamically adjusted through the sampling function.

[0045] Define the data source through the MCP roots mechanism (such as user authorization token + query conditions (such as time range)), for example, define the health data access domain as

[0046] `https: / / insurer.com / health_data`, the MCP client can also send data query requests to the MCP server, such as medical history records or Internet of Vehicles data. This interface also allows access to the hospital's HIS system.

[0047] The Roots mechanism is a function used by the client to request the server to access the resource range it needs to pay attention to. Roots is a URI (Uniform Resource Identifier) ​​used to declare and define the data access path. It can be any valid URI, including HTTP URL.

[0048] Among them, Figure 2 As shown, step S10 includes the following steps:

[0049] Step S11: upon receiving the insurance request of the target user, obtaining the basic user information and insurance type corresponding to the target user, and determining the type of information to be collected based on the basic user information and insurance type;

[0050] Step S12: Based on the type of information to be collected, determining the relevant data source from a medical database, an Internet of Vehicles database, a vehicle information database, or a body information database;

[0051] Step S13: Declare the data access boundary of each of the related data sources based on the roots mechanism, and standardize the data access path of each of the related data sources through the roots mechanism to access multiple heterogeneous data sources.

[0052] In this embodiment, the insurance types include health insurance in the medical and health field and auto insurance in the financial business field. Different types of insurance are connected to different data sources. For example, health insurance performs risk assessment based on medical records in the medical database and physical status parameters collected by wearable devices, and auto insurance performs risk assessment based on user driving behavior data and vehicle status information in the vehicle network database.

[0053] Specifically, based on the user's basic information as an identifier, insurance-related information is obtained from the corresponding data source. Then, based on the roots declaration (such as "https: / / insurer.com / health_data"), available services are filtered, and the data source path and processing rules are declared using MCP's JSONSchema. Then, a gRPC-based protocol adapter is used to achieve unified access to multimodal data, such as:

[0054] Connect to the hospital HIS system through the medical server " / get_medical_history", provide the interface / get_medical_history, and convert the XML data medical records of the hospital HIS system into JSON format through the HL7 FHIR standard converter;

[0055] Connect to the OBD device through the Internet of Vehicles server " / stream_obd_data" to transmit data in real time and expose the ` / calculate_driving_score` tool.

[0056] Step S20: Dynamically constructing context for each modal data corresponding to the insurance request from various relevant data sources to obtain insurance-related information of the target user;

[0057] In this embodiment, taking health insurance and car insurance as examples, insurance-related information is obtained respectively.

[0058] Among them, Figure 3 As shown, the insurance type includes health insurance and car insurance. When the insurance type is health insurance, step S20 includes the following steps:

[0059] Step S21: extracting medical records and physical status parameters from the medical database and the physical information database respectively according to the standardized data access path corresponding to the relevant data source;

[0060] Step S22: extracting the medical text features and physical status data features corresponding to the medical records and physical status parameters respectively, and performing cross-modal alignment and multimodal data fusion on the medical text features and physical status data features to generate the insurance-related information.

[0061] Furthermore, when the insurance type is car insurance, step S20 includes the following steps:

[0062] Extracting user driving behavior data and user vehicle information from the Internet of Vehicles database and the vehicle information database respectively based on the standardized data access path;

[0063] The insurance-related information is generated based on the user's vehicle information, the proportion of nighttime driving in the user's driving behavior data, and the number of sudden brakes.

[0064] Specifically, when the insurance type is health insurance, the medical database is accessed through the standardized path https: / / mcp-medical / api / records, and structured data (such as ICD-10 diagnosis codes, surgical records) and unstructured data (such as clinical note text) are obtained from the medical database. Physical status parameters are synchronized and obtained in real time through the wearable device API (such as "wearable: / / heart_rate"): dynamic heart rate variability (HRV), sleep quality index (PSQI) and 24-hour blood pressure fluctuation curve as insurance-related information.

[0065] Further cross-modal feature fusion of medical records and physical status parameters, including:

[0066] Medical text feature extraction involves using the BioBERT model to parse text from clinical medical records, such as text_features = biobert.encode("Patient's fasting blood glucose > 7 mmol / L in the past three months"). Next, physical data feature alignment is performed, including temporal alignment (matching blood glucose monitoring data with medication record timelines) and semantic alignment (mapping "resting heart rate > 100 bpm" to the clinical term "tachycardia"). Finally, multimodal fusion is performed to generate a fused health vector, such as health_vector = [Diabetes risk: 0.87, Cardiovascular risk: 0.42, Real-time risk fluctuation: ↑].

[0067] When the policy type is auto insurance, the system collects real-time OBD data streams through the "obd: / / accelerometer" interface: braking_events = detect_hard_brake(accel_data, threshold = 0.5g). The system also calculates the nighttime driving percentage (nighttime driving percentage = nighttime mileage / total mileage * 100%) and the frequency of hard turns (calculated based on the OBD brake pressure threshold) from the PBD data.

[0068] Vehicle status information includes vehicle age (if the vehicle age is >10 years, the risk weight increases by 15%), safety rating (if the safety rating is <3 stars, the risk weight increases by 25%), and electric / fuel vehicle type (if the vehicle is electric, the risk weight decreases by 5%).

[0069] As can be seen, in the above solution, through the Context Builder component of MCP, cross-modal attention fusion is performed based on medical records for health insurance and physical status parameters collected in real time by wearable devices, and comprehensive risk assessment is performed based on driving behavior and vehicle status information for auto insurance, significantly improving the dynamics and accuracy of risk assessment.

[0070] Step S30: Underwrite the insurance-related information through the MCP tool chain to generate a risk assessment result corresponding to the target user.

[0071] In this embodiment, various underwriting rule tools are encapsulated into an MCP tool chain, such as BMI threshold determination and driving behavior risk scoring. Using the underwriting rules and multimodal fusion features corresponding to each insurance type, the target user's insurance-related information is underwritten.

[0072] Among them, Figure 4 As shown, when the insurance type is health insurance, step S30 includes the following steps:

[0073] Step S31: Obtain, through the health insurance risk assessment model in the tool chain, a first feature distribution corresponding to the enhanced health fusion feature data and a second feature distribution corresponding to the real health sample, and calculate a distribution difference value between the first feature distribution and the second feature distribution;

[0074] Step S32: performing conflict detection on the insurance-related information and the real health sample based on the KL divergence and the distribution difference value, and determining the risk assessment result based on the conflict detection result.

[0075] Furthermore, step S30 further includes the following steps:

[0076] Calculate the target user's driving risk score based on the dynamic driving risk score formula, the proportion of nighttime driving, and the number of sudden braking times;

[0077] Based on the target weight distribution, the driving risk score and the user vehicle information, a risk level corresponding to the target user is calculated, and the risk assessment result is determined based on the risk level.

[0078] In this embodiment, the health insurance underwriting rules (conflict detection engine) in the toolchain perform distribution conflict detection on the first feature distribution corresponding to the enhanced health fusion feature data and the second feature distribution corresponding to the real health samples. For example, if a person claims to have no smoking history but the respiratory sensor detects CO > 10 ppm, or HRV is continuously < 30 ms and no heart disease has been reported, the risk of fraud is determined to be high.

[0079] Among them, the final premium = basic premium × (1 + 2 × KL divergence) × regional factor.

[0080] Through the auto insurance underwriting rules in the tool chain (such as driving risk scoring model),

[0081] Driving risk score = 100-25×√(night risk)-10×min(aggressive driving index, 5).

[0082] Based on the driving risk score, a risk assessment level is determined.

[0083] Furthermore, the risk assessment method further includes:

[0084] Calculate the premium quotation corresponding to the target user through the tool chain and the risk assessment result;

[0085] Based on the premium quotation, generate an electronic insurance policy corresponding to the insurance request, and calculate a hash value corresponding to a keyword in the electronic insurance policy;

[0086] The electronic insurance policy and its hash value are stored through the MCP blockchain server, and the corresponding evidence certificate of the electronic insurance policy is returned.

[0087] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0088] Through the above-mentioned approach, this embodiment provides a risk assessment method based on the Model Context Protocol. When receiving an insurance request from a target user, the data access boundary of at least one relevant data source is defined through the roots mechanism of the Model Context Protocol (MCP) to access multiple heterogeneous data sources. Dynamic context is constructed for each modal data corresponding to the insurance request from each relevant data source to obtain the insurance-related information of the target user. The insurance-related information is underwritten through the MCP tool chain to generate a risk assessment result corresponding to the target user. In the present invention, the roots mechanism of MCP and the pre-built tool chain are used to unify the interfaces of multi-source heterogeneous data, thereby reducing the workload of repeated development and shortening the system docking cycle. By supporting cross-system data acquisition, dynamic data aggregation is achieved, data silos are eliminated, the underwriting error rate is reduced, and the underwriting accuracy of the financial insurance system is improved. The underwriting rules are encapsulated through the MCP tool chain to achieve personalized risk assessment, thereby improving the flexibility and scalability of the financial insurance system.

[0089] In one embodiment, a risk assessment device based on a model context protocol is provided, and the risk assessment device based on a model context protocol corresponds one-to-one to the risk assessment processing method based on a model context protocol in the above embodiment. Figure 5 As shown, the risk assessment device based on the model context protocol includes a multi-source data access module 101, a related information construction module 102 and an insurance risk assessment module 103. The functional modules are described in detail as follows:

[0090] The multi-source data access module 101 is used to define the data access boundary of at least one relevant data source through the roots mechanism of the Model Context Protocol (MCP) upon receiving the insurance application request of the target user, so as to access multiple heterogeneous data sources;

[0091] The relevant information construction module 102 is used to dynamically construct contexts for each modal data corresponding to the insurance request from various relevant data sources to obtain insurance-related information of the target user;

[0092] The insurance risk assessment module 103 is used to underwrite the insurance-related information through the MCP tool chain and generate a risk assessment result corresponding to the target user.

[0093] The multi-source data access module 101 is further configured to:

[0094] Upon receiving the insurance request from the target user, obtaining the basic user information and insurance type corresponding to the target user, and determining the type of information to be collected based on the basic user information and insurance type;

[0095] Based on the type of information to be collected, determining the relevant data source from a medical database, an Internet of Vehicles database, a vehicle information database, or a physical information database;

[0096] The data access boundary of each of the related data sources is declared based on the roots mechanism, and the data access path of each of the related data sources is standardized through the roots mechanism to access multiple heterogeneous data sources.

[0097] Furthermore, the insurance type includes health insurance and car insurance. When the insurance type is health insurance, the relevant information construction module 102 is further used to:

[0098] extracting medical records and physical status parameters from the medical database and the physical information database respectively based on the standardized data access path corresponding to the relevant data source;

[0099] The medical text features and physical status data features corresponding to the medical records and physical status parameters are extracted respectively, and the medical text features and physical status data features are cross-modally aligned and multi-modally fused to generate the insurance-related information.

[0100] Furthermore, when the insurance type is health insurance, the insurance risk assessment module is further configured to:

[0101] Obtaining, through the health insurance risk assessment model in the tool chain, a first feature distribution corresponding to the enhanced health fusion feature data and a second feature distribution corresponding to the real health sample, and calculating a distribution difference value between the first feature distribution and the second feature distribution;

[0102] Based on the KL divergence and the distribution difference value, conflict detection is performed on the insurance-related information and the real health sample, and the risk assessment result is determined based on the conflict detection result.

[0103] Furthermore, when the insurance type is car insurance, the relevant information building module 102 is further used to:

[0104] Extracting user driving behavior data and user vehicle information from the Internet of Vehicles database and the vehicle information database respectively based on the standardized data access path;

[0105] The insurance-related information is generated based on the user's vehicle information, the proportion of nighttime driving in the user's driving behavior data, and the number of sudden brakes.

[0106] Furthermore, when the insurance type is auto insurance, the insurance risk assessment module 103 is further configured to:

[0107] Calculate the target user's driving risk score based on the dynamic driving risk score formula, the proportion of nighttime driving, and the number of sudden braking times;

[0108] Based on the target weight distribution, the driving risk score and the user vehicle information, a risk level corresponding to the target user is calculated, and the risk assessment result is determined based on the risk level.

[0109] Furthermore, the policy storage module of the risk assessment device based on the model context protocol is used to:

[0110] Calculate the premium quotation corresponding to the target user through the tool chain and the risk assessment result;

[0111] Based on the premium quotation, generate an electronic insurance policy corresponding to the insurance request, and calculate a hash value corresponding to a keyword in the electronic insurance policy;

[0112] The electronic insurance policy and its hash value are stored through the MCP blockchain server, and the corresponding evidence certificate of the electronic insurance policy is returned.

[0113] For the specific definition of the risk assessment device based on the model context protocol, please refer to the definition of the risk assessment method based on the model context protocol above, which will not be repeated here. The various modules in the above-mentioned risk assessment device based on the model context protocol can be implemented in whole or in part by software, hardware and their combination. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0114] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 6 As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile and / or volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external client via a network connection. When the computer program is executed by the processor, it implements the functions or steps of a risk assessment method based on a model context protocol.

[0115] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the following steps are performed:

[0116] When receiving the insurance request from the target user, the data access boundary of at least one relevant data source is defined through the roots mechanism of the Model Context Protocol (MCP) to access multiple heterogeneous data sources;

[0117] Dynamically constructing context for each modal data corresponding to the insurance request from various relevant data sources to obtain insurance-related information for the target user;

[0118] The insurance-related information is underwritten through the MCP tool chain to generate a risk assessment result corresponding to the target user.

[0119] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0120] Upon receiving the insurance request from the target user, obtaining the basic user information and insurance type corresponding to the target user, and determining the type of information to be collected based on the basic user information and insurance type;

[0121] Based on the type of information to be collected, determining the relevant data source from a medical database, an Internet of Vehicles database, a vehicle information database, or a physical information database;

[0122] The data access boundary of each of the related data sources is declared based on the roots mechanism, and the data access path of each of the related data sources is standardized through the roots mechanism to access multiple heterogeneous data sources.

[0123] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0124] When the insurance type is health insurance, extracting medical records and physical status parameters from the medical database and the physical information database respectively based on the standardized data access path corresponding to the relevant data source;

[0125] The medical text features and physical status data features corresponding to the medical records and physical status parameters are extracted respectively, and the medical text features and physical status data features are cross-modally aligned and multi-modally fused to generate the insurance-related information.

[0126] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0127] Obtaining, through the health insurance risk assessment model in the tool chain, a first feature distribution corresponding to the enhanced health fusion feature data and a second feature distribution corresponding to the real health sample, and calculating a distribution difference value between the first feature distribution and the second feature distribution;

[0128] Based on the KL divergence and the distribution difference value, conflict detection is performed on the insurance-related information and the real health sample, and the risk assessment result is determined based on the conflict detection result.

[0129] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0130] When the insurance type is auto insurance, extracting user driving behavior data and user vehicle information from the Internet of Vehicles database and the vehicle information database respectively based on the standardized data access path;

[0131] The insurance-related information is generated based on the user's vehicle information, the proportion of nighttime driving in the user's driving behavior data, and the number of sudden brakes.

[0132] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0133] Calculate the target user's driving risk score based on the dynamic driving risk score formula, the proportion of nighttime driving, and the number of sudden braking times;

[0134] Based on the target weight distribution, the driving risk score and the user vehicle information, a risk level corresponding to the target user is calculated, and the risk assessment result is determined based on the risk level.

[0135] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0136] Calculate the premium quotation corresponding to the target user through the tool chain and the risk assessment result;

[0137] Based on the premium quotation, generate an electronic insurance policy corresponding to the insurance request, and calculate a hash value corresponding to a keyword in the electronic insurance policy;

[0138] The electronic insurance policy and its hash value are stored through the MCP blockchain server, and the corresponding evidence certificate of the electronic insurance policy is returned.

[0139] It should be noted that the above functions or steps that can be implemented by the computer-readable storage medium or computer device can refer to the relevant description in the aforementioned method embodiment. To avoid repetition, they will not be described one by one here.

[0140] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0141] Those skilled in the art will clearly understand that for the sake of convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0142] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention, and should all be included in the scope of protection of the present invention.

Claims

1. A risk assessment method based on model context protocol, characterized in that: The method comprises: When receiving the insurance request from the target user, the data access boundary of at least one relevant data source is defined through the roots mechanism of the Model Context Protocol (MCP) to access multiple heterogeneous data sources; Dynamically constructing context for each modal data corresponding to the insurance request from various relevant data sources to obtain insurance-related information for the target user; The insurance-related information is underwritten through the MCP tool chain to generate a risk assessment result corresponding to the target user.

2. The risk assessment method based on model context protocol according to claim 1, characterized in that: When receiving the insurance application request of the target user, defining the data access boundary of at least one relevant data source through the roots mechanism of the Model Context Protocol (MCP) to access multiple heterogeneous data sources, including: Upon receiving the insurance request from the target user, obtaining the basic user information and insurance type corresponding to the target user, and determining the type of information to be collected based on the basic user information and insurance type; Based on the type of information to be collected, determining the relevant data source from a medical database, an Internet of Vehicles database, a vehicle information database, or a physical information database; The data access boundary of each of the related data sources is declared based on the roots mechanism, and the data access path of each of the related data sources is standardized through the roots mechanism to access multiple heterogeneous data sources.

3. The risk assessment method based on model context protocol according to claim 2, characterized in that: The insurance types include health insurance and car insurance. Dynamic context construction is performed on each modal data corresponding to the insurance request from various relevant data sources to obtain insurance-related information of the target user, including: When the insurance type is health insurance, extracting medical records and physical status parameters from the medical database and the physical information database respectively based on the standardized data access path corresponding to the relevant data source; The medical text features and physical status data features corresponding to the medical records and physical status parameters are extracted respectively, and the medical text features and physical status data features are cross-modally aligned and multi-modally fused to generate the insurance-related information.

4. The risk assessment method based on model context protocol according to claim 3, characterized in that: The MCP tool chain is used to underwrite the insurance-related information and generate a risk assessment result corresponding to the target user, including: Obtaining, through the health insurance risk assessment model in the tool chain, a first feature distribution corresponding to the enhanced health fusion feature data and a second feature distribution corresponding to the real health sample, and calculating a distribution difference value between the first feature distribution and the second feature distribution; Based on the KL divergence and the distribution difference value, conflict detection is performed on the insurance-related information and the real health sample, and the risk assessment result is determined based on the conflict detection result.

5. The risk assessment method based on model context protocol according to claim 3, characterized in that: The dynamically constructing contexts for each modal data corresponding to the insurance request from each relevant data source to obtain insurance-related information of the target user further includes: When the insurance type is auto insurance, extracting user driving behavior data and user vehicle information from the Internet of Vehicles database and the vehicle information database respectively based on the standardized data access path; The insurance-related information is generated based on the user's vehicle information, the proportion of nighttime driving in the user's driving behavior data, and the number of sudden brakes.

6. The risk assessment method based on model context protocol according to claim 5, characterized in that: The MCP tool chain is used to underwrite the insurance-related information and generate a risk assessment result corresponding to the target user, including: Calculate the target user's driving risk score based on the dynamic driving risk score formula, the proportion of nighttime driving, and the number of sudden braking times; Based on the target weight distribution, the driving risk score and the user vehicle information, a risk level corresponding to the target user is calculated, and the risk assessment result is determined based on the risk level.

7. The risk assessment method based on model context protocol according to any one of claims 1 to 6, characterized in that: The risk assessment method further comprises: Calculate the premium quotation corresponding to the target user through the tool chain and the risk assessment result; Based on the premium quotation, generate an electronic insurance policy corresponding to the insurance request, and calculate a hash value corresponding to a keyword in the electronic insurance policy; The electronic insurance policy and its hash value are stored through the MCP blockchain server, and the corresponding evidence certificate of the electronic insurance policy is returned.

8. A risk assessment device based on a model context protocol, characterized in that: The device comprises: The multi-source data access module is used to define the data access boundary of at least one relevant data source through the roots mechanism of the Model Context Protocol (MCP) upon receiving the insurance application request of the target user, so as to access multiple heterogeneous data sources; A related information construction module is used to dynamically construct context for each modal data corresponding to the insurance request from various relevant data sources to obtain insurance related information of the target user; The insurance risk assessment module is used to underwrite the insurance-related information through the MCP tool chain and generate a risk assessment result corresponding to the target user.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the risk assessment method based on the model context protocol according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the risk assessment method based on the model context protocol according to any one of claims 1 to 7 are implemented.

Images