Insurance pricing method, device, electronic device and medium based on data fusion

Abstract

The embodiments of the present application provide an insurance pricing method, device, electronic device, and medium based on data fusion, which belongs to the field of financial technology. The method includes: in response to a user's pricing instruction, obtaining the basic product information corresponding to the pricing instruction through the pricing system; receiving the identity information sent by the user through the pricing system, and determining the user-related data and the credit platform bound to the pricing system based on the identity information; in response to the user's authorization instruction, obtaining the user's corresponding credit score through the credit platform; determining the comprehensive discount parameters based on the credit score and the user-related data; performing a risk assessment on the basic product information, credit score, and user-related data to obtain a risk assessment result; performing a pricing analysis on the risk assessment result and the comprehensive discount parameters through a pricing model to obtain the target insurance pricing. The embodiments of the present application fully reflect the customer's risk level through multi-dimensional data, thereby realizing dynamic adjustment of the customer's insurance pricing.

Description

Technical Field

[0001] The present application relates to the field of financial technology, and in particular to an insurance pricing method, device, electronic device and medium based on data fusion. Background Art

[0002] The rapid development of drone technology has driven the growth of the low-altitude economy, with drones playing a vital role in a variety of industries, including logistics, agriculture, and infrastructure inspection. However, due to the numerous risk factors associated with drone operations, such as pilot operating habits, weather conditions, and terrain complexity, existing insurance pricing systems often rely on long-term, fixed pricing models that fail to flexibly reflect customers' individual needs. This rigidity not only limits the development of personalized insurance products but also ignores the actual impact of customer behavior on risk levels. This prevents insurance companies from fully leveraging comprehensive customer data to formulate more reasonable insurance pricing, thereby failing to fully reflect customers' true risk levels. This results in customers being unable to enjoy dynamic premium discounts at different stages of their use, reducing their insurance purchasing experience and retention. Summary of the Invention

[0003] The main purpose of the embodiments of the present application is to propose an insurance pricing method, device, electronic device and medium based on data fusion, which can fully reflect the customer's risk level through multi-dimensional data and realize dynamic adjustment of the customer's insurance pricing.

[0004] To achieve the above objectives, a first aspect of an embodiment of the present application proposes an insurance pricing method based on data fusion, the method comprising:

[0005] In response to a user's pricing instruction, obtaining, through a pricing system, basic product information corresponding to the pricing instruction;

[0006] Receiving, through the pricing system, identity information sent by the user, and determining, based on the identity information, user-related data and a credit platform bound to the pricing system, wherein the user-related data is used to represent the user's purchase record and claim information;

[0007] In response to the user's authorization instruction, obtaining the user's corresponding credit score through the credit platform;

[0008] determining a comprehensive discount parameter based on the credit score and the user association data;

[0009] Performing a risk assessment on the product basic information, the credit score, and the user-related data to obtain a risk assessment result;

[0010] The risk assessment results and the comprehensive discount parameters are subjected to pricing analysis using a preset pricing model to obtain target insurance pricing.

[0011] In some embodiments, determining a comprehensive discount parameter based on the credit score and the user association data includes:

[0012] Determining the highest credit score and the lowest credit score of the user within a preset period of time based on the credit score;

[0013] Normalizing the credit score according to the highest credit score and the lowest credit score to obtain a target credit score;

[0014] performing an exponential decay operation on the target credit score based on a preset adjustment parameter to obtain a credit score discount, wherein the adjustment parameter is used to control the decay rate of the discount;

[0015] Determine the number of claims, insurance type weights, and total number of policies based on the user-related data;

[0016] Determine a relevance discount based on the number of claims, the insurance type weight, and the total number of policies;

[0017] A comprehensive discount parameter is determined according to the credit score discount and the relevance discount.

[0018] In some embodiments, determining a comprehensive discount parameter based on the credit score discount and the relevance discount includes:

[0019] determining a first priority corresponding to the credit score discount, and determining a second priority corresponding to the relevance discount;

[0020] Setting a first weight parameter corresponding to the credit score discount according to the first priority, and setting a second weight parameter corresponding to the relevance discount according to the second priority;

[0021] multiplying the credit score discount by the first weight parameter to obtain a credit score, and multiplying the relevance discount by the second weight parameter to obtain a relevance score;

[0022] A comprehensive discount parameter is determined based on the credit score and the relevance score.

[0023] In some embodiments, performing risk assessment on the product basic information, the credit score, and the user-related data to obtain a risk assessment result includes:

[0024] Determining a product identifier and product history data based on the product basic information, and performing a correlation analysis on the credit score and the user association data based on the product identifier to obtain a correlation relationship;

[0025] Taking the product identification as a node, and connecting a plurality of the nodes according to the association relationship to construct a network graph;

[0026] Performing feature fusion on the network graph through a preset graph neural network to obtain feature representation;

[0027] Performing feature processing on the product historical data to obtain time series features and non-time series features;

[0028] Performing data prediction on the time series features through a preset long short-term memory network to obtain a hidden state;

[0029] The hidden state is feature enhanced based on a preset attention mechanism to obtain comprehensive information, and the risk score of the time series feature is calculated based on the comprehensive information and the non-time series feature to obtain a risk assessment result.

[0030] In some embodiments, the performing feature fusion on the network graph through a preset graph neural network to obtain a feature representation includes:

[0031] For each node in the network graph, assign a feature vector to the node;

[0032] Selecting a test node in the network graph and determining an initial feature vector of the test node and a set of neighboring nodes around the test node;

[0033] Performing a linear transformation on the neighbor nodes in the neighbor node set by the graph neural network to obtain a first parameter, and performing a linear transformation on the test node to obtain a second parameter;

[0034] Determining a set of neighbor feature vectors corresponding to the set of neighbor nodes, determining a first feature representation based on the initial feature vector, and determining a second feature representation based on the set of neighbor feature vectors;

[0035] The first parameter, the second parameter, the first feature representation, and the second feature representation are input into the graph neural network, so that the graph neural network performs a hierarchical update on the test node based on the first parameter, the second parameter, the first feature representation, and the second feature representation to obtain a feature representation.

[0036] In some embodiments, the hidden state is enhanced based on a preset attention mechanism to obtain comprehensive information, and a risk score of the time series feature is calculated based on the comprehensive information and the non-time series feature to obtain a risk assessment result, including:

[0037] Calculating an attention score of the hidden state based on a preset attention mechanism;

[0038] Determine the attention weight of the current time step according to the attention score;

[0039] Performing weighted summation on the context of the time series feature based on the attention weight and the hidden state to obtain comprehensive information;

[0040] Splicing the comprehensive information and the non-time series features to obtain splicing features;

[0041] Perform risk scoring on the splicing features to obtain a risk assessment result.

[0042] In some embodiments, performing pricing analysis on the risk assessment results and the comprehensive discount parameters using a preset pricing model to obtain target insurance pricing includes:

[0043] Determine the basic insurance premium for the user based on the user-related data;

[0044] A pricing analysis is performed based on the basic premium, the risk assessment results, and the comprehensive discount parameters to obtain target insurance pricing.

[0045] To achieve the above-mentioned objectives, a second aspect of an embodiment of the present application provides an insurance pricing device based on data fusion; the device includes:

[0046] A data acquisition module, configured to respond to a user's pricing instruction and acquire, through a pricing system, basic product information corresponding to the pricing instruction;

[0047] a data determination module, configured to receive identity information sent by the user through the pricing system, and determine user-related data and a credit platform bound to the pricing system based on the identity information, wherein the user-related data is used to represent the user's purchase history and claim information;

[0048] a credit score acquisition module, configured to obtain the credit score corresponding to the user through the credit platform in response to an authorization instruction from the user;

[0049] a discount calculation module, configured to determine a comprehensive discount parameter based on the credit score and the user association data;

[0050] A risk assessment module, configured to perform a risk assessment on the product basic information, the credit score, and the user-related data to obtain a risk assessment result;

[0051] The pricing analysis module is used to perform pricing analysis on the risk assessment results and the comprehensive discount parameters through a preset pricing model to obtain target insurance pricing.

[0052] To achieve the above-mentioned purpose, the third aspect of an embodiment of the present application proposes an electronic device, an electronic device comprising a memory and a processor, the memory storing a computer program, and the processor implementing the insurance pricing method based on data fusion as described in the first aspect when executing the computer program.

[0053] To achieve the above-mentioned purpose, the fourth aspect of an embodiment of the present application proposes a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the insurance pricing method based on data fusion as described in the first aspect.

[0054] The insurance pricing method, device, electronic device and storage medium based on data fusion proposed in this application respond to the user's pricing instructions, obtain the basic product information corresponding to the pricing instructions through the pricing system, thereby obtaining the basic information of the insurance product, which is convenient for subsequent use as a reference for risk assessment. The pricing system then receives the identity information sent by the user, and determines the user-related data and the credit platform bound to the pricing system based on the identity information. It can obtain the user's purchase records of other products and the current claims records, so as to determine whether the user has made a claim and the specific information of the claim. In response to the user's authorization instruction, the credit score corresponding to the user is obtained through the credit platform, so that the credit situation of the user can be reflected through the credit score, the accuracy of risk management can be improved, and risk losses can be reduced. Then, the comprehensive discount parameters are determined based on the credit score and user-related data, so that the discount parameters can be reasonably set according to the user's credit situation, historical claims situation and insurance purchase situation. After that, a risk assessment is performed on the basic information of the product, the credit score and the user-related data. By analyzing the multi-dimensional data, the risk level of the customer is fully reflected, and the risk assessment result is obtained to achieve a comprehensive assessment of the insurance risk. Then, the risk assessment result and the comprehensive discount parameter are priced and analyzed by the preset pricing model. The premium is flexibly adjusted in combination with the user's insurance situation and actual risk to obtain the target insurance pricing, so that more personalized premium discounts can be provided to the user. The embodiment of the present application dynamically adjusts the user's insurance premium rate through multi-dimensional data collection, fusion and analysis, and can provide users with more personalized premium discounts by connecting to a third-party credit platform and integrating data such as purchase records of other insurance products. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 This is a flow chart of an insurance pricing method based on data fusion provided in an embodiment of the present application;

[0056] Figure 2 yes Figure 1 Flowchart of step S104 in FIG.

[0057] Figure 3 yes Figure 2 Flowchart of step S206 in FIG.

[0058] Figure 4 yes Figure 1 Flowchart of step S105 in FIG.

[0059] Figure 5 yes Figure 4 Flowchart of step S403 in FIG.

[0060] Figure 6 yes Figure 4 Flowchart of step S406 in FIG.

[0061] Figure 7 yes Figure 1 Flowchart of step S106 in FIG.

[0062] Figure 8 This is a schematic diagram of the structure of an insurance pricing device based on data fusion provided in an embodiment of the present application;

[0063] Figure 9 This is a schematic diagram of the hardware structure of the electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0064] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0065] It should be noted that although the device schematics illustrate functional module divisions and the flowcharts illustrate logical sequences, in certain circumstances, the steps shown or described may be performed in a sequence that differs from the module divisions in the device or the sequence in the flowcharts. The terms "first," "second," and so on, in the specification, claims, and drawings, are used to distinguish similar items and are not necessarily used to describe a specific sequence or precedence.

[0066] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application pertains. The terms used herein are for the purpose of describing the embodiments of this application only and are not intended to limit this application.

[0067] First, let’s analyze some of the terms used in this application:

[0068] Natural Language Processing (NLP): NLP uses computers to process, understand, and apply human languages ​​(such as Chinese and English). NLP is a branch of artificial intelligence and an interdisciplinary subject between computer science and linguistics. It is often referred to as computational linguistics. Natural language processing includes grammatical analysis, semantic analysis, and text understanding. Natural language processing is commonly used in technical fields such as machine translation, handwritten and printed character recognition, speech recognition and text-to-speech conversion, information intent recognition, information extraction and filtering, text classification and clustering, public opinion analysis, and opinion mining. It involves data mining related to language processing, machine learning, knowledge acquisition, knowledge engineering, artificial intelligence research, and linguistic research related to language computing.

[0069] The data fusion-based insurance pricing method and device, electronic device, and storage medium provided in the embodiments of the present application can fully reflect the customer's risk level through multi-dimensional data and achieve dynamic adjustment of the customer's insurance pricing.

[0070] The following examples are used to illustrate the details. First, the insurance pricing method based on data fusion in the embodiments of the present application is described.

[0071] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, techniques, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results.

[0072] Fundamental AI technologies generally include sensors, specialized AI chips, cloud computing, distributed storage, big data processing, operating / interaction systems, and mechatronics. AI software technologies primarily encompass computer vision, robotics, biometrics, module management for online customer service systems, natural language processing, and machine learning / deep learning.

[0073] The insurance pricing method based on data fusion provided in the embodiment of the present application relates to the field of financial technology. The insurance pricing method based on data fusion provided in the embodiment of the present application can be applied to a terminal, can be applied to a server side, or can be software running in a terminal or a server side. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.; the server side can be configured as an independent physical server, or as a server cluster or distributed system composed of multiple physical servers, or as a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application that implements an insurance pricing method based on data fusion, etc., but is not limited to the above forms.

[0074] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld devices or portable devices, tablet devices, multi-processor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data information types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0075] It should be noted that in each specific embodiment of the present application, when it comes to the need to perform relevant processing based on data related to the user's identity or characteristics, such as user information, user behavior data, user historical data, and user location information, the user's permission or consent will be obtained first, and the collection, use, and processing of such data will comply with the relevant laws, regulations, and standards of the relevant countries and regions. In addition, when the embodiment of the present application needs to obtain the user's sensitive personal information, the user's separate permission or consent will be obtained through a pop-up window or by jumping to a confirmation page. After clearly obtaining the user's separate permission or consent, the necessary user-related data for the normal operation of the embodiment of the present application will be obtained.

[0076] The rapid development of drone technology has driven the growth of the low-altitude economy, with drones playing a vital role in a variety of industries, including logistics, agriculture, and infrastructure inspection. However, due to the numerous risk factors associated with drone operations, such as pilot operating habits, weather conditions, and terrain complexity, existing insurance pricing systems often rely on long-term, fixed pricing models that fail to flexibly reflect customers' individual needs. This rigidity not only limits the development of personalized insurance products but also ignores the actual impact of customer behavior on risk levels. This prevents insurance companies from fully leveraging comprehensive customer data to formulate more reasonable insurance pricing, thereby failing to fully reflect customers' true risk levels. This results in customers being unable to enjoy dynamic premium discounts at different stages of their use, reducing their insurance purchasing experience and retention.

[0077] In order to solve the above problems, this embodiment provides an insurance pricing method, device, electronic device and storage medium based on data fusion. In response to the user's pricing instruction, the pricing system obtains the basic product information corresponding to the pricing instruction, thereby obtaining the basic information of the insurance product, which is convenient for subsequent use as a reference for risk assessment. The pricing system then receives the identity information sent by the user, and determines the user-related data and the credit platform bound to the pricing system based on the identity information. The user's purchase records of other products and current claims records can be obtained, so that it can be determined whether the user has made a claim and the specific information of the claim. In response to the user's authorization instruction, the credit score corresponding to the user is obtained through the credit platform, so that the credit situation of the user can be reflected through the credit score, the accuracy of risk management can be improved, and risk losses can be reduced. Then, the comprehensive discount parameters are determined based on the credit score and user-related data, so that the discount parameters can be reasonably set according to the user's credit situation, historical claims situation and insurance purchase situation. After that, a risk assessment is performed on the basic information of the product, the credit score and the user-related data. By analyzing the multi-dimensional data, the risk level of the customer is fully reflected, and the risk assessment result is obtained to achieve a comprehensive assessment of the insurance risk. Then, the risk assessment result and the comprehensive discount parameter are priced and analyzed by the preset pricing model. The premium is flexibly adjusted in combination with the user's insurance situation and actual risk to obtain the target insurance pricing, so that more personalized premium discounts can be provided to the user. The embodiment of the present application dynamically adjusts the user's insurance premium rate through multi-dimensional data collection, fusion and analysis, and can provide users with more personalized premium discounts by connecting to a third-party credit platform and integrating data such as purchase records of other insurance products.

[0078] The following is a detailed description with reference to the accompanying drawings.

[0079] Figure 1 This is an optional flow chart of the insurance pricing method based on data fusion provided in the embodiment of the present application. Figure 1The method may include but is not limited to steps S101 to S106.

[0080] Step S101: In response to a user's pricing instruction, basic product information corresponding to the pricing instruction is obtained through a pricing system.

[0081] In step S101 of some embodiments, the user can send a pricing instruction by triggering the relevant control on the pricing system. In response to the user's pricing instruction, the pricing system obtains the basic product information corresponding to the pricing instruction, wherein the basic product information includes the basic information of the product, such as brand, model, market pricing, etc., which is convenient for subsequent use as a basic reference for risk assessment.

[0082] It should be noted that the pricing system in the embodiment of the present application is provided with an intelligent question-and-answer module, which includes multiple preset questions, wherein the preset questions can be single-choice questions, multiple-choice questions, or fill-in-the-blank questions. When the preset questions are single-choice questions or multiple-choice questions, the user can select the corresponding options to answer; when the preset questions are fill-in-the-blank questions, the user can answer in the answer area corresponding to each preset question. For example, the intelligent question-and-answer module is provided with questions such as flight frequency, flight habits, and degree of knowledge of drones, and provides corresponding options. At this time, the user can select the corresponding options for different questions to answer. After the user answers, the corresponding pricing instructions are generated to carry out the insurance pricing process.

[0083] Step S102: The pricing system receives the identity information sent by the user, and determines the user association data and the credit platform bound to the pricing system according to the identity information.

[0084] It should be noted that user-related data is used to represent the user's purchase history and claims information.

[0085] In step S102 of some embodiments, after obtaining the basic product information, the embodiments of the present application will also receive the identity information sent by the user through the pricing system, and determine the user-related data based on the identity information, so that the user can determine other insurance products that the user has purchased or currently holds from this insurance company through the user-related data, such as car insurance, pet insurance, health insurance, etc. At the same time, the user-related data can also be used to determine the user's previous claims history, so as to facilitate the subsequent dynamic pricing based on the user's purchase history and claims history, thereby improving the flexibility of insurance pricing. At the same time, the embodiments of the present application will also determine the credit platform bound to the pricing system based on the identity information, such as the Alipay platform, WeChat platform, etc., to facilitate the subsequent acquisition of the user's credit score.

[0086] Step S103: In response to the user's authorization instruction, the credit score corresponding to the user is obtained through the credit platform.

[0087] In step S103 of some embodiments, in response to the user's authorization instruction, the pricing system sends a query request instruction to the credit platform. The credit platform queries the credit score corresponding to the user's identity information and returns the corresponding result. The embodiment of the present application obtains the credit score corresponding to the user through the credit platform, thereby obtaining the user's credit status, which facilitates the subsequent risk assessment of the premium.

[0088] It is understandable that the credit score in the embodiment of the present application can be determined according to the type of credit platform. For example, when the credit platform is Alipay, the credit score at this time is the Alipay Sesame Credit score; when the credit platform is WeChat, the credit score at this time is the WeChat Credit score, etc. The embodiment of the present application does not make specific restrictions.

[0089] It is worth noting that the pricing system in the embodiment of the present application will update the credit score in real time to ensure the accuracy of the assessment, wherein the higher the credit score, the greater the premium discount the user will receive.

[0090] Step S104: Determine comprehensive discount parameters based on the credit score and user association data.

[0091] In step S104 of some embodiments, a comprehensive discount parameter is determined based on the credit score and user-related data, so that people with higher credit scores can obtain more obvious discounts and obtain credit score discounts. At the same time, it can avoid the sharp changes caused by the logarithmic function at low scores, realize smooth calculation of user premium discounts, ensure that the discount differences between different credit scores are not too significant, and balance the discounts for customers with low credit scores and high credit scores.

[0092] Step S105: Perform risk assessment on the product basic information, credit score and user-related data to obtain a risk assessment result.

[0093] In step S105 of some embodiments, a risk assessment is performed on the basic product information, credit score, and user-related data to obtain a risk assessment result, thereby enabling a comprehensive risk assessment to be achieved by combining multi-dimensional information, improving the accuracy of the risk assessment, and achieving a comprehensive analysis of historical data.

[0094] Step S106: Perform pricing analysis on the risk assessment results and the comprehensive discount parameters using a preset pricing model to obtain target insurance pricing.

[0095] In step S106 of some embodiments, a pricing analysis is performed on the risk assessment results and the comprehensive discount parameters through a preset pricing model, so that a comprehensive risk assessment can be achieved by combining multi-dimensional information, target insurance pricing can be obtained, the accuracy of risk assessment can be improved, and a comprehensive analysis of historical data can be achieved.

[0096] In some embodiments, the present invention also uses blockchain technology to record data, ensuring transparency and immutability during each premium calculation and adjustment process. Data is also asynchronously uploaded to the blockchain. Specifically, considering the impact of blockchain performance, data is queued as text, files, or images and uploaded to the blockchain within 5 minutes to ensure efficiency.

[0097] Specifically, data from each premium calculation and adjustment process is signed and uploaded to the blockchain. Data uploaded to the blockchain can be in the form of text, files, or images, ensuring immutability and traceability at every stage of premium adjustments. To address the limitations of on-chain operations, the system utilizes an asynchronous upload mechanism, queuing data for upload, ensuring efficient data processing. Furthermore, the system utilizes Bayesian optimization to implement real-time pricing adjustments based on the on-chain data, combined with data from third-party platforms and internal historical data, ensuring personalized and fair premiums for customers.

[0098] In some embodiments, the present invention can also flexibly adjust premiums based on a user's real-time operational behavior, user-related data, and credit score, ensuring that premiums align with the user's actual risk level. Users with different mission types, operating environments, and credit scores will receive different premium adjustment policies. For example, if a drone user has performed multiple high-risk missions but has a high credit score, the system will grant an additional 5% premium discount based on their credit history.

[0099] See also Figure 2 In some embodiments, step S104 may also include but is not limited to steps S201 to S206.

[0100] Step S201: Determine the highest credit score and the lowest credit score of the user within a preset time period based on the credit score.

[0101] Step S202 : normalize the credit scores according to the highest credit score and the lowest credit score to obtain a target credit score.

[0102] Step S203: performing an exponential decay operation on the target credit score based on preset adjustment parameters to obtain a credit score discount.

[0103] It should be noted that the adjustment parameter is used to control the decay rate of the discount.

[0104] In steps S201 to S203 of some embodiments, in the process of determining the comprehensive discount parameters based on the credit score and user-related data, the embodiments of the present application first determine the highest credit score and the lowest credit score of the user within a preset time period based on the credit score. By determining the lowest and highest credit scores, a standardized comparison benchmark can be provided for the credit status of different users or entities, facilitating the subsequent precise adjustment of premium pricing. The credit score is then normalized based on the highest credit score and the lowest credit score so that the value ranges of different features are consistent, thereby accelerating the convergence process of optimization algorithms such as gradient descent to obtain a target credit score. Thereafter, an exponential decay operation is performed on the target credit score based on the preset adjustment parameters, so that higher credit scores can obtain more obvious discounts, thereby obtaining a credit score discount. At the same time, the sharp changes caused by the logarithmic function at low scores can be avoided, and smooth calculation of the user's premium discount can be achieved, ensuring that the discount differences between different credit scores are not too significant, and balancing the discounts for customers with low credit scores and high credit scores.

[0105] Specifically, in the embodiment of the present application, the exponential decay operation on the target credit score is expressed as follows:

[0106] ;

[0107] in, is the normalized credit score, i.e., the target credit score; is the highest credit score; is the minimum credit score; It is the original credit score.

[0108] ;

[0109] in, is the normalized credit score; is a tuning parameter used to control the discount decay rate, It’s a credit score discount.

[0110] It is understandable that in the embodiment of the present application, customers with higher credit scores enjoy greater discounts, while ensuring that the discounts for customers with lower credit scores are not too low. For example, the discount gap between a customer with a credit score of 700 and a customer with a credit score of 800 is smaller.

[0111] Step S204: Determine the number of claims, insurance type weights, and the total number of policies based on the user-related data.

[0112] Step S205: Determine the correlation discount based on the number of claims, the insurance type weight, and the total number of policies.

[0113] Step S206: determining a comprehensive discount parameter based on the credit score discount and the relevance discount.

[0114] In steps S204 to S206 of some embodiments, the embodiments of the present application will also determine the number of claims, insurance type weights and the total number of policies based on user-related data, thereby achieving comprehensive collection of the user's historical claims data and purchased insurance data, and then determine the correlation discount based on the number of claims, insurance type weights and the total number of policies, so that users who have purchased more products or have fewer claims have higher corresponding premiums, thereby achieving flexible adjustment of premium pricing. Afterwards, comprehensive discount parameters are determined based on credit score discounts and correlation discounts, and dynamic adjustment of discount parameters is achieved through comprehensive combination of multiple data. Discount parameters can be determined based on the user's product purchase situation, historical claims situation and credit score situation at the same time, thereby improving the accuracy and flexibility of insurance pricing.

[0115] Specifically, the relevance discount in the embodiment of the present application is calculated as follows:

[0116] ;

[0117] in, is the relevance discount; is the insurance type weight; is the number of claims; is the total number of policies.

[0118] It can be understood that in the embodiment of the present application, the insurance types for which the user customers have no claims records will receive the maximum weighted discount, and the insurance types with a higher number of claims will have a reduced corresponding discount ratio.

[0119] It should be noted that the preset duration in the embodiment of the present application can be set according to the user's needs, for example, one year, two years or half a month, etc., and the embodiment of the present application does not impose any specific restrictions.

[0120] See also Figure 3 In some embodiments, step S206 may also include but is not limited to steps S301 to S304.

[0121] Step S301 : determining a first priority corresponding to a credit score discount, and determining a second priority corresponding to a relevance discount.

[0122] Step S302 : setting a first weight parameter corresponding to the credit score discount according to the first priority, and setting a second weight parameter corresponding to the relevance discount according to the second priority.

[0123] Step S303: multiply the credit score discount by the first weight parameter to obtain a credit score, and multiply the relevance discount by the second weight parameter to obtain a relevance score.

[0124] Step S304: Determine a comprehensive discount parameter based on the credit score and the relevance score.

[0125] In steps S301 to S304 of some embodiments, in the process of determining the comprehensive discount parameters based on the credit score discount and the relevance discount, the embodiments of the present application first determine the first priority corresponding to the credit score discount, and determine the second priority corresponding to the relevance discount, thereby determining the priority relationship between the credit score discount and the relevance discount, and then set the first weight parameter corresponding to the credit score discount according to the first priority, and set the second weight parameter corresponding to the relevance discount according to the second priority, so that different weight parameters can be set for different priorities, and a larger weight parameter can be set for a discount with a higher priority, and a smaller weight parameter can be set for a discount with a lower priority. Afterwards, the credit score discount is multiplied by the first weight parameter to obtain a credit score value, and the relevance discount is multiplied by the second weight parameter to obtain a relevance score. Then, the comprehensive discount parameters are determined based on the credit score value and the relevance score, so that the comprehensive discount parameters can be reasonably determined, which is convenient for providing personalized services and products to users in the future.

[0126] Specifically, the operation of determining the comprehensive discount parameters in the embodiment of the present application is as follows:

[0127] ;

[0128] in, discounts for credit scores; Discount for relevance; is the first weight parameter; is the second weight parameter.

[0129] It can be understood that the sum of the weights of the first weight parameter and the second weight parameter in the embodiment of the present application is 1, and the first priority and the second priority can be set according to the needs of the user. For example, the first priority is set higher than the second priority. At this time, the first weight parameter is greater than the second weight parameter. The embodiment of the present application does not impose specific restrictions.

[0130] See also Figure 4 In some embodiments, step S105 may also include but is not limited to steps S401 to S406.

[0131] Step S401: determine the product identification and product history data based on the basic product information, and perform correlation analysis on the credit score and user association data based on the product identification to obtain a correlation relationship.

[0132] Step S402: Use product identifiers as nodes and connect multiple nodes according to association relationships to construct a network graph.

[0133] Step S403: Perform feature fusion on the network graph through a preset graph neural network to obtain feature representation.

[0134] Step S404: perform feature processing on the product history data to obtain time series features and non-time series features.

[0135] Step S405: Data prediction is performed on the time series features through a preset long short-term memory network to obtain a hidden state.

[0136] In step S406, the hidden state is enhanced based on the preset attention mechanism to obtain comprehensive information, and the risk score of the time series feature is calculated based on the comprehensive information and the non-time series feature to obtain a risk assessment result.

[0137] In some embodiments, in steps S401 to S406, during the process of risk assessment of product basic information, credit score, and user-related data, the embodiment of the present application first determines the product identification and product historical data based on the product basic information, and performs correlation analysis on the credit score and user-related data based on the product identification, analyzes the correlation between the credit score and user-related data and the product identification, obtains the correlation relationship, and then uses the product identification as a node, and connects multiple nodes according to the correlation relationship to construct a network graph, wherein each node represents a different product, and the edge between the nodes represents the correlation between different nodes. The edge between the two nodes is used as a common feature of the data source, so that the correspondence between the products can be clearly described, and the relationship between the product data can be intuitively displayed through the graph. Afterwards, the network graph is subjected to feature fusion through a preset graph neural network, which can improve the model convergence speed while ensuring the balance of feature fusion and obtain feature representation. Then, feature processing is performed on the product historical data, and the product historical data is converted into time series features. Normalization and standardization are used to ensure that the feature values ​​are within a reasonable range, and time series features and non-time series features are obtained. After that, data prediction is performed on the time series features through the preset long short-term memory network, so that the long short-term memory network can capture the risk pattern of the product and pay attention to the key risk points in the risk pattern to obtain the hidden state. Finally, the hidden state is feature enhanced based on the preset attention mechanism, so that comprehensive information of important moments in the time series can be obtained, and the risk score of the time series feature is calculated based on the comprehensive information and non-time series features to obtain the risk assessment result, so that a comprehensive risk assessment can be achieved by combining multi-dimensional information, the accuracy of risk assessment can be improved, and a comprehensive analysis of historical data can be achieved.

[0138] Specifically, the specific process of performing data prediction on time series features using a preset long short-term memory network in the embodiment of the present application is as follows:

[0139] = ;

[0140] = ;

[0141] ;

[0142] ;

[0143] in, is the weight matrix of the forget gate, which is used to control the forgetting ratio of the previous state. is the weight matrix of the input gate, which determines the extent to which new input data enters the cell state. It is the fusion matrix of input data and previous state, used to update the cell state. Represents the forget gate, which controls the state information of the previous time step. It is the input gate that controls the inflow of new information. It is the control signal of the output gate, which is used to determine the weight of the output at the current moment. For cell states, long-term dependencies in the time series are recorded. is the hidden state output at the current moment. The sigmoid function is used to compress the value between 0 and 1. The input for the current time step. is the bias term of the forget gate. is the bias term of the input gate. is the bias term of the cell state. is the input for the current time step.

[0144] See also Figure 5 In some embodiments, step S403 may also include but is not limited to steps S501 to S505.

[0145] Step S501: For each node in the network graph, a feature vector is assigned to the node.

[0146] Step S502: Select a test node in the network graph, and determine the initial feature vector of the test node and a set of neighbor nodes around the test node.

[0147] Step S503: Perform a linear transformation on the neighbor nodes in the neighbor node set through a graph neural network to obtain a first parameter, and perform a linear transformation on the test node to obtain a second parameter.

[0148] Step S504 : determining a set of neighbor feature vectors corresponding to the set of neighbor nodes, determining a first feature representation based on the initial feature vector, and determining a second feature representation based on the set of neighbor feature vectors.

[0149] Step S505: Input the first parameter, the second parameter, the first feature representation, and the second feature representation into the graph neural network, so that the graph neural network hierarchically updates the test node based on the first parameter, the second parameter, the first feature representation, and the second feature representation to obtain a feature representation.

[0150] In some embodiments, in steps S501 to S505, in the process of performing feature fusion on the network graph through a preset graph neural network, for each node in the network graph, a feature vector is assigned to the node, and the feature vector is used as the initial feature vector, wherein these feature vectors can be attributes of the node, for example, the brand and model of the drone, etc., and then a test node is randomly selected in the network graph, and the initial feature vector of the test node and the set of neighbor nodes around the test node are determined, thereby realizing message transmission between the test node and the neighbor nodes, and then the neighbor nodes in the neighbor node set are linearly transformed by the graph neural network to obtain the first parameter, and the test node is initialized and initialized. The test node is linearly transformed to obtain the second parameter. Then, the set of neighbor feature vectors corresponding to the set of neighbor nodes is determined, and the first feature representation is determined based on the initial feature vector, and the second feature representation is determined based on the set of neighbor feature vectors. Finally, the first parameter, the second parameter, the first feature representation and the second feature representation are input into the graph neural network, so that the graph neural network performs a hierarchical update on the test node based on the first parameter, the second parameter, the first feature representation and the second feature representation, which can improve the convergence speed of the model while ensuring the balance of feature fusion, and obtain the feature representation of the information of the aggregated neighbor nodes and their own information, so that features can be extracted in multidimensional data and the accuracy of risk assessment can be improved.

[0151] Specifically, the embodiment of the present application performs a hierarchical update on the test node as follows:

[0152] ;

[0153] in, Indicates the Test nodes in the layer Feature representation; Represents a test node The set of neighbor nodes of Represents the normalization coefficient, which is used to prevent the node eigenvalue from exploding or disappearing, ensure the balance of feature fusion, and improve the model convergence speed; represents the first feature representation; represents the second feature representation; represents the weight matrix; represents a nonlinear activation function; Representation and test nodes Adjacent neighbor nodes.

[0154] See also Figure 6 In some embodiments, step S406 may also include but is not limited to steps S601 to S605.

[0155] Step S601: Calculate the attention score of the hidden state based on the preset attention mechanism.

[0156] Step S602: Determine the attention weight of the current time step based on the attention score.

[0157] Step S603: Perform weighted summation on the context of the time series features based on the attention weight and the hidden state to obtain comprehensive information.

[0158] Step S604: splicing the comprehensive information and the non-time series features to obtain splicing features.

[0159] Step S605: Perform risk scoring on the splicing features to obtain a risk assessment result.

[0160] In steps S601 to S605 of some embodiments, in the process of performing feature enhancement on the hidden state based on a preset attention mechanism to obtain comprehensive information, and calculating the risk score of the time series feature based on the comprehensive information and the non-time series features to obtain the risk assessment result, the embodiment of the present application first calculates the attention score of the hidden state based on the preset attention mechanism, so that the importance of the current moment can be measured by the attention score, and then determines the attention weight of the current time step based on the attention score, so that the relative importance of the current time step can be measured. Thereafter, the context of the time series feature is weightedly summed based on the attention weight and the hidden state, so as to obtain comprehensive information of important moments in the time series, and then the comprehensive information and the non-time series features are spliced ​​to obtain the spliced ​​features. Thereafter, the spliced ​​features are risk scored to obtain the risk assessment result, so that a comprehensive risk assessment can be achieved by combining multi-dimensional information, the accuracy of the risk assessment can be improved, and a comprehensive analysis of historical data can be achieved.

[0161] Specifically, the calculation process of weighted summation of the context of time series features based on attention weights and hidden states is as follows:

[0162] ;

[0163] ;

[0164] ;

[0165] in, is the weight matrix used to calculate the attention score. is the bias term of the attention mechanism. is the time step An attention score that measures the importance of the current moment. Time step The attention weight indicates the relative importance of the moment. The context vector after weighted summation represents the comprehensive information of important moments in the time series.

[0166] Specifically, the calculation process of the risk score of the splicing features in the embodiment of the present application is as follows:

[0167] ;

[0168] in, ] represents the time series context vector and non-time series features splicing features. Represents the weight matrix, used to connect comprehensive features To risk score . Represents the partial rank vector, used to Adjust the linear output of the model together. Represents an activation function, such as the sigmoid function.

[0169] See also Figure 7 In some embodiments, step S106 may also include but is not limited to steps S701 to S702.

[0170] Step S701: Determine the basic insurance premium of the user according to the user association data.

[0171] Step S702: Perform pricing analysis based on the basic premium, risk assessment results, and comprehensive discount parameters to obtain target insurance pricing.

[0172] In steps S701 to S702 of some embodiments, in the process of performing pricing analysis on the risk assessment results and comprehensive discount parameters through a preset pricing model, the embodiment of the present application first determines the basic premium of the user's insurance based on the user-related data, and then performs pricing analysis based on the basic premium, risk assessment results and comprehensive discount parameters to obtain the target insurance pricing. It can dynamically adjust the user's insurance premium rate in combination with multi-dimensional data to achieve more personalized pricing.

[0173] Specifically, the embodiment of the present application uses Bayesian optimization technology to dynamically adjust the customer's insurance premium rate. The specific process is as follows:

[0174] ;

[0175] in, time Personalized premiums, namely target insurance pricing. As basic premium. is the risk adjustment factor. is the comprehensive discount parameter.

[0176] ;

[0177] in, is the mean function of the forecast distribution. is the covariance function, which is used to measure the similarity between data points.

[0178] See also Figure 8 The embodiment of the present application further provides an insurance pricing device based on data fusion, the device comprising:

[0179] The data acquisition module 801 is used to respond to the user's pricing instruction and obtain the basic product information corresponding to the pricing instruction through the pricing system;

[0180] Data determination module 802, configured to receive identity information sent by a user through the pricing system and determine user-related data and a credit platform bound to the pricing system based on the identity information, wherein the user-related data is used to represent the user's purchase history and claim information;

[0181] The credit score acquisition module 803 is used to obtain the user's corresponding credit score through the credit platform in response to the user's authorization instruction;

[0182] Discount calculation module 804, used to determine comprehensive discount parameters based on credit scores and user association data;

[0183] The risk assessment module 805 is used to perform risk assessment on basic product information, credit scores, and user-related data to obtain risk assessment results;

[0184] The pricing analysis module 806 is used to perform pricing analysis on the risk assessment results and comprehensive discount parameters through a preset pricing model to obtain target insurance pricing.

[0185] The specific implementation of the insurance pricing device based on data fusion is basically the same as the specific embodiment of the insurance pricing method based on data fusion described above, and will not be repeated here.

[0186] An embodiment of the present application further provides an electronic device comprising: a memory, a processor, a program stored in the memory and executable on the processor, and a data bus for enabling communication between the processor and the memory. When the program is executed by the processor, the aforementioned insurance pricing method based on data fusion is implemented. The electronic device may be any intelligent terminal, including a tablet computer and an in-vehicle computer.

[0187] See also Figure 9 , Figure 9 The hardware structure of an electronic device according to another embodiment is shown. The electronic device includes:

[0188] The processor 901 may be implemented as a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is configured to execute relevant programs to implement the technical solutions provided in the embodiments of the present application.

[0189] Memory 902 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). Memory 902 can store an operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in memory 902 and is called by processor 901 to execute the insurance pricing method based on data fusion in the embodiments of this application.

[0190] Input / output interface 903, used to implement information input and output;

[0191] Communication interface 904, used to implement communication interaction between this device and other devices, which can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WiFi, Bluetooth, etc.);

[0192] Bus 905 , which transmits information between various components of the device (e.g., processor 901 , memory 902 , input / output interface 903 , and communication interface 904 );

[0193] The processor 901 , the memory 902 , the input / output interface 903 and the communication interface 904 are connected to each other in communication within the device via a bus 905 .

[0194] An embodiment of the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the above-mentioned insurance pricing method based on data fusion.

[0195] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely arranged relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0196] The data fusion-based insurance pricing method, device, electronic device, and storage medium provided in the embodiments of the present application respond to a user's pricing instruction and obtain the basic product information corresponding to the pricing instruction through the pricing system, thereby obtaining the basic information of the insurance product, which is convenient for subsequent use as a reference for risk assessment. The pricing system then receives the identity information sent by the user, and determines the user-related data and the credit platform bound to the pricing system based on the identity information. The user's purchase records of other products and the current claims record can be obtained, so that it can be determined whether the user has made a claim and the specific information of the claim. In response to the user's authorization instruction, the credit score corresponding to the user is obtained through the credit platform, so that the credit situation of the user can be reflected through the credit score, the accuracy of risk management can be improved, and risk losses can be reduced. Then, the comprehensive discount parameters are determined based on the credit score and user-related data, so that the discount parameters can be reasonably set according to the user's credit situation, historical claims situation and insurance purchase situation. After that, a risk assessment is performed on the basic information of the product, the credit score and the user-related data. By analyzing the multi-dimensional data, the risk level of the customer is fully reflected, and the risk assessment result is obtained to achieve a comprehensive assessment of the insurance risk. Then, the risk assessment result and the comprehensive discount parameter are priced and analyzed by the preset pricing model. The premium is flexibly adjusted in combination with the user's insurance situation and actual risk to obtain the target insurance pricing, so that more personalized premium discounts can be provided to the user. The embodiment of the present application dynamically adjusts the user's insurance premium rate through multi-dimensional data collection, fusion and analysis, and can provide users with more personalized premium discounts by connecting to a third-party credit platform and integrating data such as purchase records of other insurance products.

[0197] The embodiments described in the embodiments of this application are intended to more clearly illustrate the technical solutions of the embodiments of this application and do not constitute a limitation on the technical solutions provided by the embodiments of this application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

[0198] It will be understood by those skilled in the art that Figure 1-9 The technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than those shown in the figures, or a combination of certain steps, or different steps.

[0199] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, i.e., they may be located in one place or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of this embodiment.

[0200] Those skilled in the art will appreciate that all or some of the steps in the methods, systems, and functional modules / units in the devices disclosed above may be implemented as software, firmware, hardware, or appropriate combinations thereof.

[0201] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0202] It should be understood that in this application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following items" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0203] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the above-mentioned units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0204] The units described above as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0205] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0206] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes multiple instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present application. The aforementioned storage medium includes: various media that can store programs, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0207] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but are not intended to limit the scope of the present invention. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and essence of the present invention should be within the scope of the present invention.

Claims

1. An insurance pricing method based on data fusion, characterized in that: The method comprises: In response to a user's pricing instruction, obtaining, through a pricing system, basic product information corresponding to the pricing instruction; Receiving, through the pricing system, identity information sent by the user, and determining, based on the identity information, user-related data and a credit platform bound to the pricing system, wherein the user-related data is used to represent the user's purchase record and claim information; In response to the user's authorization instruction, obtaining the user's corresponding credit score through the credit platform; determining a comprehensive discount parameter based on the credit score and the user-related data; Performing a risk assessment on the basic product information, the credit score, and the user-related data to obtain a risk assessment result; performing a pricing analysis on the risk assessment result and the comprehensive discount parameter using a preset pricing model to obtain a target insurance price; The determining of the comprehensive discount parameter according to the credit score and the user association data includes: Determining the highest credit score and the lowest credit score of the user within a preset time period based on the credit score; normalizing the credit score based on the highest credit score and the lowest credit score to obtain a target credit score; performing an exponential decay operation on the target credit score based on a preset adjustment parameter to obtain a credit score discount, wherein the adjustment parameter is used to control the decay rate of the discount; determining the number of claims, the insurance type weight, and the total number of policies based on the user-related data; determining a relevance discount based on the number of claims, the insurance type weight, and the total number of policies; and determining a comprehensive discount parameter based on the credit score discount and the relevance discount; The risk assessment of the product basic information, the credit score, and the user-related data to obtain a risk assessment result includes: Determine the product identification and product historical data based on the product basic information, and perform correlation analysis on the credit score and the user-related data based on the product identification to obtain a correlation relationship; use the product identification as a node, and connect multiple nodes based on the correlation relationship to construct a network graph; perform feature fusion on the network graph through a preset graph neural network to obtain a feature representation; perform feature processing on the product historical data to obtain time series features and non-time series features; perform data prediction on the time series features through a preset long short-term memory network to obtain a hidden state; perform feature enhancement on the hidden state based on a preset attention mechanism to obtain comprehensive information, and calculate the risk score of the time series features based on the comprehensive information and the non-time series features to obtain a risk assessment result.

2. The insurance pricing method based on data fusion according to claim 1, characterized in that: The determining of a comprehensive discount parameter according to the credit score discount and the relevance discount includes: determining a first priority corresponding to the credit score discount, and determining a second priority corresponding to the relevance discount; Setting a first weight parameter corresponding to the credit score discount according to the first priority, and setting a second weight parameter corresponding to the relevance discount according to the second priority; multiplying the credit score discount by the first weight parameter to obtain a credit score, and multiplying the relevance discount by the second weight parameter to obtain a relevance score; A comprehensive discount parameter is determined based on the credit score and the relevance score.

3. The insurance pricing method based on data fusion according to claim 1, characterized in that: The feature fusion of the network graph by a preset graph neural network to obtain a feature representation includes: For each node in the network graph, assign a feature vector to the node; Selecting a test node in the network graph and determining an initial feature vector of the test node and a set of neighboring nodes around the test node; Performing a linear transformation on the neighbor nodes in the neighbor node set by the graph neural network to obtain a first parameter, and performing a linear transformation on the test node to obtain a second parameter; Determining a set of neighbor feature vectors corresponding to the set of neighbor nodes, determining a first feature representation based on the initial feature vector, and determining a second feature representation based on the set of neighbor feature vectors; The first parameter, the second parameter, the first feature representation, and the second feature representation are input into the graph neural network, so that the graph neural network performs a hierarchical update on the test node based on the first parameter, the second parameter, the first feature representation, and the second feature representation to obtain a feature representation.

4. The insurance pricing method based on data fusion according to claim 1, characterized in that: The hidden state is enhanced based on the preset attention mechanism to obtain comprehensive information, and the risk score of the time series feature is calculated based on the comprehensive information and the non-time series feature to obtain a risk assessment result, including: Calculating an attention score of the hidden state based on a preset attention mechanism; Determine the attention weight of the current time step according to the attention score; Performing weighted summation on the context of the time series feature based on the attention weight and the hidden state to obtain comprehensive information; Splicing the comprehensive information and the non-time series features to obtain splicing features; Perform risk scoring on the splicing features to obtain a risk assessment result.

5. The insurance pricing method based on data fusion according to claim 1, characterized in that: The risk assessment results and the comprehensive discount parameters are subjected to pricing analysis using a preset pricing model to obtain target insurance pricing, including: Determine the basic insurance premium for the user based on the user-related data; A pricing analysis is performed based on the basic premium, the risk assessment results, and the comprehensive discount parameters to obtain target insurance pricing.

6. An insurance pricing device based on data fusion, characterized in that: The device comprises: A data acquisition module, configured to respond to a user's pricing instruction and acquire, through a pricing system, basic product information corresponding to the pricing instruction; a data determination module, configured to receive identity information sent by the user through the pricing system, and determine user-related data and a credit platform bound to the pricing system based on the identity information, wherein the user-related data is used to represent the user's purchase history and claim information; a credit score acquisition module, configured to obtain the credit score corresponding to the user through the credit platform in response to an authorization instruction from the user; a discount calculation module, configured to determine a comprehensive discount parameter based on the credit score and the user association data; A risk assessment module, configured to perform a risk assessment on the product basic information, the credit score, and the user-related data to obtain a risk assessment result; A pricing analysis module, configured to perform pricing analysis on the risk assessment results and the comprehensive discount parameters using a preset pricing model to obtain a target insurance price; The determining of the comprehensive discount parameter according to the credit score and the user association data includes: Determining the highest credit score and the lowest credit score of the user within a preset time period based on the credit score; normalizing the credit score based on the highest credit score and the lowest credit score to obtain a target credit score; performing an exponential decay operation on the target credit score based on a preset adjustment parameter to obtain a credit score discount, wherein the adjustment parameter is used to control the decay rate of the discount; determining the number of claims, the insurance type weight, and the total number of policies based on the user-related data; determining a relevance discount based on the number of claims, the insurance type weight, and the total number of policies; and determining a comprehensive discount parameter based on the credit score discount and the relevance discount; The risk assessment of the product basic information, the credit score, and the user-related data to obtain a risk assessment result includes: Determine the product identification and product historical data based on the product basic information, and perform correlation analysis on the credit score and the user-related data based on the product identification to obtain a correlation relationship; use the product identification as a node, and connect multiple nodes based on the correlation relationship to construct a network graph; perform feature fusion on the network graph through a preset graph neural network to obtain a feature representation; perform feature processing on the product historical data to obtain time series features and non-time series features; perform data prediction on the time series features through a preset long short-term memory network to obtain a hidden state; perform feature enhancement on the hidden state based on a preset attention mechanism to obtain comprehensive information, and calculate the risk score of the time series features based on the comprehensive information and the non-time series features to obtain a risk assessment result.

7. An electronic device, characterized in that: The electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the insurance pricing method based on data fusion as described in any one of claims 1 to 5 when executing the computer program.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the insurance pricing method based on data fusion according to any one of claims 1 to 5 is implemented.

Images