Underwriting method, device and equipment based on artificial intelligence and storage medium

Abstract

The embodiment of the invention relates to the field of artificial intelligence, and provides an artificial intelligence-based underwriting method and device, a server and a storage medium, and the method comprises the steps: obtaining the underwriting related data of a to-be-underwritten target customer; determining an initial underwriting result of the target customer based on the underwriting related data through a preset target underwriting model; acquiring medical image data and target physiological data of the target customer in response to the initial underwriting result that underwriting is passed; determining a health risk assessment result of the target customer based on the medical image data through a preset health risk assessment model; and according to the health risk assessment result and the target physiological data, verifying the initial underwriting result, and according to a verification result of the initial underwriting result, determining a target underwriting result of the target customer. The embodiment of the invention can be applied to businesses such as financial science and technology, medical health, old-age care and the like so as to improve the underwriting efficiency and accuracy of insurance products such as life insurance new contracts and the like.

Description

Technical Field

[0001] The present application relates to the technical field of intelligent decision-making, and in particular to an underwriting method, apparatus, device and storage medium based on artificial intelligence. Background Art

[0002] After a customer purchases a new life insurance policy, the insurance company must conduct underwriting. Currently, underwriting for new life insurance policies and other insurance products relies on manual review and historical underwriting records. Insurance company staff review documents such as the customer's medical records and refer to similar underwriting cases. However, this method is time-consuming and lacks assurance of accuracy, leading to a higher risk of claims. Therefore, improving underwriting efficiency and accuracy is a pressing issue. Summary of the Invention

[0003] The embodiments of the present application provide an artificial intelligence-based underwriting method, apparatus, device, and storage medium, aiming to improve the efficiency and accuracy of underwriting.

[0004] In a first aspect, an embodiment of the present application provides an artificial intelligence-based underwriting method, applied to a first server, the method comprising:

[0005] Obtain underwriting-related data of target customers to be underwritten;

[0006] determining an initial underwriting result for the target customer based on the underwriting-related data using a preset target underwriting model, wherein the target underwriting model is a pre-trained machine learning model;

[0007] In response to the initial underwriting result being an approved underwriting, obtaining medical image data and target physiological data of the target customer;

[0008] Determining a health risk assessment result of the target customer based on the medical image data using a preset health risk assessment model, wherein the health risk assessment model is a pre-trained deep learning model;

[0009] The initial underwriting result is verified based on the health risk assessment result and the target physiological data, and the target underwriting result of the target customer is determined based on the verification result of the initial underwriting result.

[0010] In a second aspect, an embodiment of the present application further provides an artificial intelligence-based underwriting device, the underwriting device comprising:

[0011] The acquisition module is used to obtain underwriting-related data of the target customers to be underwritten;

[0012] an underwriting module, configured to determine an initial underwriting result for the target customer based on the underwriting-related data using a preset target underwriting model, wherein the target underwriting model is a pre-trained machine learning model;

[0013] The acquisition module is further configured to acquire the medical image data and target physiological data of the target customer in response to the initial underwriting result being an approved underwriting;

[0014] A health risk assessment module, configured to determine a health risk assessment result of the target customer based on the medical image data using a preset health risk assessment model, wherein the health risk assessment model is a pre-trained deep learning model;

[0015] a verification module, configured to verify the initial underwriting result based on the health risk assessment result and the target physiological data;

[0016] A determination module is used to determine the target underwriting result of the target customer based on the verification result of the initial underwriting result.

[0017] In a third aspect, an embodiment of the present application further provides a server, comprising a processor, a memory, and a computer program stored on the memory and executable by the processor, wherein when the computer program is executed by the processor, the steps of the artificial intelligence-based underwriting method as described in the first aspect are implemented.

[0018] In a fourth aspect, an embodiment of the present application further provides a storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the steps of the artificial intelligence-based underwriting method as described in the first aspect are implemented.

[0019] The embodiments of the present application provide an underwriting method, apparatus, device and storage medium based on artificial intelligence. The embodiments of the present application use a pre-trained machine learning model as a target underwriting model to automatically determine the initial underwriting result of the target customer based on the underwriting-related data of the target customer. Then, when the initial underwriting result is underwriting passed, a pre-trained deep learning model is used as a health risk assessment model to automatically determine the health risk assessment result of the target customer based on the medical image data of the target customer. The initial underwriting result is further verified based on the health risk assessment result and the target physiological data. Finally, according to the verification result, the target underwriting result of the target customer is determined. The entire underwriting process does not require the participation of insurance company staff, reduces the impact of human factors on underwriting, improves the efficiency and accuracy of underwriting, and the underwriting result is further verified based on the health risk assessment result and the target physiological data, further improving the accuracy of underwriting, thereby reducing the risk of insurance company claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0021] Figure 1 This is a schematic diagram of a scenario for implementing the artificial intelligence-based underwriting method provided in an embodiment of the present application;

[0022] Figure 2 This is another schematic diagram of a scenario for implementing the artificial intelligence-based underwriting method provided in an embodiment of the present application;

[0023] Figure 3 This is a flowchart of the steps of an artificial intelligence-based underwriting method provided in an embodiment of the present application;

[0024] Figure 4 This is a schematic diagram of the steps of another artificial intelligence-based underwriting method provided in an embodiment of the present application;

[0025] Figure 5 This is a schematic diagram of the steps of another artificial intelligence-based underwriting method provided in an embodiment of the present application;

[0026] Figure 6 This is a schematic block diagram of an artificial intelligence-based underwriting device provided in an embodiment of the present application;

[0027] Figure 7 This is a schematic block diagram of the structure of a server provided in an embodiment of the present application.

[0028] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0029] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0030] The flowcharts shown in the accompanying drawings are for illustrative purposes only and do not necessarily include all contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may be decomposed, combined, or partially merged, so the actual execution order may vary depending on the actual situation.

[0031] After a customer purchases an insurance product, such as a new life insurance contract, the insurance company needs to underwrite the customer. Currently, underwriting for new life insurance contracts and other insurance products relies on manual review and historical underwriting data. Insurance company staff must review the customer's medical records and other documents to determine the customer's health risks and then determine whether the underwriting is approved. However, this method is time-consuming and cannot guarantee the accuracy of the underwriting, which increases the risk of insurance companies paying claims. Therefore, improving the efficiency and accuracy of underwriting is a pressing issue.

[0032] To solve the above problems, the embodiments of the present application provide an underwriting method, apparatus, device and storage medium based on artificial intelligence. The embodiments of the present application use a pre-trained machine learning model as the target underwriting model to automatically determine the initial underwriting result of the target customer based on the underwriting-related data of the target customer. Then, when the initial underwriting result is underwriting passed, the pre-trained deep learning model is used as the health risk assessment model to automatically determine the health risk assessment result of the target customer based on the medical image data of the target customer. The initial underwriting result is further verified based on the health risk assessment result and the target physiological data. Finally, based on the verification result, the target underwriting result of the target customer is determined. The entire underwriting process does not require the participation of insurance company staff, reduces the impact of human factors on underwriting, improves the efficiency and accuracy of underwriting, and the underwriting result is further verified based on the health risk assessment result and the target physiological data, further improving the accuracy of underwriting, thereby reducing the risk of insurance company claims.

[0033] 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.

[0034] See also Figure 1 , Figure 1 It is a schematic diagram of a scenario for implementing the artificial intelligence-based underwriting method provided in an embodiment of the present application.

[0035] like Figure 1As shown, the scenario includes a first server 10 and a second server 20, and the first server 10 is communicatively connected to the second server 20. The first server 10 stores a target underwriting model and a health risk assessment model. The target underwriting model is a pre-trained machine learning model, and the health risk assessment model is a pre-trained deep learning model. The second server 20 is used to integrate medical image data from multiple medical institutions and multiple health care institutions, so that the first server 10 can obtain the customer's medical image data in real time through the second server 20. The second server 20 is also used to store the customer's physiological data, which is collected by the wearable device worn by the customer and uploaded to the second server 20. The wearable device may include a smart bracelet, a smart watch, a smart ring, a smart armband, and smart glasses.

[0036] In some embodiments, the first server 10 obtains underwriting-related data of a target customer to be underwritten; the first server 10 determines an initial underwriting result of the target customer based on the underwriting-related data through a preset target underwriting model; in response to the initial underwriting result being a passed underwriting, the first server 10 obtains medical image data of the target customer through the second server 20, and obtains target physiological data from the physiological database of the target customer stored in the second server 20, wherein the difference between the acquisition time of the target physiological data and the current time is less than or equal to the preset difference, and the physiological data in the physiological database is collected through a wearable device worn by the target customer; the first server 10 determines a health risk assessment result of the target customer based on the medical image data through a preset health risk assessment model; the first server 10 verifies the initial underwriting result according to the health risk assessment result and the target physiological data, and determines the target underwriting result of the target customer based on the verification result of the initial underwriting result.

[0037] The above embodiment constructs a collaborative architecture of the first server 10 (intelligent underwriting engine) and the second server 20 (medical data middle platform), solves the problem of isolated islands of health data (medical data and physiological data), and realizes real-time linkage underwriting of health data across institutions, which not only improves the efficiency of obtaining medical data, but also avoids the risk of tampering of information submitted by customers themselves. In addition, through the underwriting closed-loop mechanism of initial underwriting → in-depth analysis of medical images → verification of initial underwriting results based on health risks and real-time physiological data, not only can the customer's organic lesions be identified to solve the pain point of missed underwriting due to hidden diseases of customers, but also the customer's real-time physiological data can be monitored to solve the pain point of insufficient data timeliness, thereby reducing the underwriting misjudgment rate and improving the accuracy of underwriting results, which can significantly reduce the insurance company's payment risk.

[0038] See also Figure 2 , Figure 2It is a schematic diagram of another scenario for implementing the artificial intelligence-based underwriting method provided in an embodiment of the present application.

[0039] like Figure 2 As shown, the scenario includes a first server 10, a second server 20, and a third server 30. The first server 10 is in communication with the second server 20 and the third server 30, respectively. The first server 10 stores a target underwriting model and a health risk assessment model. The target underwriting model is a pre-trained machine learning model, and the health risk assessment model is a pre-trained deep learning model. The second server 20 is used to integrate medical image data from multiple medical institutions and multiple health care institutions, so that the first server 10 can obtain the customer's medical image data in real time through the second server 20. The second server 20 is also used to store the customer's physiological data, which is collected through the wearable device worn by the customer and uploaded to the second server 20. The third server 30 includes an intelligent consultation module, through which the customer's health behavior data can be collected. The user's health behavior data includes the user's exercise habits, sleeping habits, eating habits, whether he drinks alcohol, whether he smokes, etc.

[0040] In some embodiments, the first server 10 obtains underwriting-related data of the target customer to be underwritten; the first server 10 determines the initial underwriting result of the target customer based on the underwriting-related data through a preset target underwriting model; in response to the initial underwriting result being a passed underwriting, the first server 10 obtains the medical image data of the target customer through the second server 20, and obtains the target physiological data from the physiological database of the target customer stored in the second server 20, wherein the difference between the acquisition time of the target physiological data and the current time is less than or equal to the preset difference, and the physiological data in the physiological database is collected through a wearable device worn by the target customer; the first server 10 obtains the health behavior data of the target customer from the third server 30; the first server 10 determines the health risk assessment result of the target customer based on the medical image data through a preset health risk assessment model; the first server 10 verifies the initial underwriting result according to the health risk assessment result, the target physiological data and the health behavior data, and determines the target underwriting result of the target customer based on the verification result of the initial underwriting result.

[0041] The above embodiment constructs a collaborative architecture of the first server 10 (intelligent underwriting engine), the second server 20 (medical data middle platform) and the third server 30 (intelligent consultation engine), which solves the problem of isolated islands of health data (medical data and physiological data) and realizes real-time linkage underwriting of health data across institutions. It not only improves the efficiency of obtaining medical data, but also avoids the risk of tampering of information submitted by customers themselves. It can ensure the authenticity of medical image data, physiological data and health behavior data. At the same time, through the initial underwriting → in-depth analysis of medical images → verification of the initial underwriting results based on health risks, real-time physiological data and health behavior data, the underwriting closed-loop mechanism can not only identify the customer's organic lesions to solve the pain point of missed underwriting due to the customer's hidden disease, but also monitor the customer's real-time physiological data to solve the pain point of insufficient data timeliness. At the same time, it can also quantify the customer's lifestyle risk based on health risk data to solve the pain point that lifestyle risk is difficult to quantify, thereby further significantly reducing the underwriting misjudgment rate, improving the accuracy of the underwriting results, and further significantly reducing the insurance company's payment risk.

[0042] It should be noted that the embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is a theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. The first server 10, the second server 20 and the third server 30 can be independent servers, or they can be cloud servers that provide 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, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0043] The following will be combined Figure 1 and / or Figure 2 The scenario in the embodiment of this application provides a detailed introduction to the artificial intelligence-based underwriting method. Figure 1 and / or Figure 2 The scenarios are only used to explain the artificial intelligence underwriting method provided in the embodiments of the present application, but do not constitute a limitation on the application scenarios of the artificial intelligence underwriting method provided in the embodiments of the present application.

[0044] See also Figure 3 , Figure 3 This is a flowchart of the steps of an artificial intelligence-based underwriting method provided in an embodiment of the present application.

[0045] like Figure 3 As shown, the underwriting method includes steps S101 to S105.

[0046] Step S101: Obtain underwriting-related data of the target customer to be underwritten.

[0047] In this embodiment, the target customer may be a customer who newly purchases a life insurance contract or other insurance product. The underwriting-related data refers to data related to underwriting, which may include the target customer's age, gender, occupation, health behavior data obtained by manually inquiring the target customer, financial information, health status information and insurance history information, etc. Health behavior data includes exercise habits, sleeping habits, eating habits, whether to drink alcohol, whether to smoke, etc. Financial information includes annual income, fixed assets and liabilities, etc. Health status information may include existing diseases and medical history, physical examination reports (height and weight (BMI), blood pressure, liver function, kidney function and other indicators), family medical history, medical records (previous hospitalizations, surgeries, chronic disease treatments or long-term medication).

[0048] Step S102: Determine the initial underwriting result of the target customer based on underwriting-related data through a preset target underwriting model, where the target underwriting model is a pre-trained machine learning model.

[0049] In this embodiment, the pre-trained machine learning model may include a pre-trained neural network model or a gradient boosting tree model. For example, the target underwriting model is obtained by pre-training a neural network model based on a plurality of first sample data, where the first sample data includes underwriting-related data samples and annotated first underwriting results.

[0050] Specifically, the training process of the above-mentioned neural network model may include: initializing the neural network model; selecting a first sample data from multiple first sample data as the current sample data; inputting the underwriting-related data sample in the current sample data into the neural network model for underwriting processing to obtain a predicted second underwriting result; determining the loss value based on the first underwriting result in the current sample data and the predicted second underwriting result; when the loss value is greater than the preset first loss value, adjusting the parameters of the neural network model to update the neural network model, and returning to execute the step of selecting a first sample data from multiple first sample data as the current sample data; when the loss value is less than or equal to the preset first loss value, stopping training the neural network model to obtain the target underwriting model.

[0051] Step S103: In response to the initial underwriting result being approved, obtaining medical image data and target physiological data of the target customer.

[0052] In this embodiment, the medical image data includes X-rays, computed tomography (CT) images, single photon emission computed tomography (SPECT) images, magnetic resonance imaging (MRI), functional magnetic resonance imaging (fMRI), positron emission tomography (PET) images, ultrasound images, and microscopic images. The difference between the acquisition time of the target physiological data and the current time is less than or equal to the preset difference, and the target physiological data includes heart rate, blood pressure, blood oxygen, etc. Among them, the preset difference can be set based on actual conditions, and the embodiments of the present application do not specifically limit this. For example, the preset difference includes 1 hour, 2 hours, 5 hours, 12 hours, 18 hours, or 24 hours.

[0053] In some embodiments, in response to the initial underwriting result being an underwriting pass, obtaining the target customer's medical image data and target physiological data may include: in response to the initial underwriting result being an underwriting pass, obtaining the target customer's medical image data through a second server; obtaining the target physiological data from a physiological database of the target customer stored on the second server, wherein the difference between the acquisition time of the target physiological data and the current time is less than or equal to a preset difference, and the physiological data in the physiological database is collected through a wearable device worn by the target customer.

[0054] In some embodiments, medical image data and physiological data are also stored on a blockchain. Smart contracts can be used to enable authorized access and dynamic updates to these data across institutions. The storage of medical image data and physiological data on a blockchain in this embodiment supports cross-institutional health data sharing, ensuring the security and privacy of health data, and providing highly secure data storage and transmission in compliance with medical privacy regulations.

[0055] Step S104: Determine the health risk assessment result of the target customer based on the medical image data through a preset health risk assessment model, where the health risk assessment model is a pre-trained deep learning model.

[0056] In this embodiment, the pre-trained deep learning model may include a convolutional neural network model (CNN), a transformer model, a recurrent neural network model (RNN), and a deep convolutional neural network model (DNN). For example, the health risk assessment model is obtained by pre-training the deep learning model based on a plurality of second sample data, where the second sample data includes medical image samples and annotated first health risk assessment results. The deep learning model may include a multi-scale feature extraction layer and a self-attention module.

[0057] Specifically, the training process of the above-mentioned deep learning model may include: initializing the deep learning model; selecting a second sample data from multiple second sample data as the current sample data; inputting the medical image sample in the current sample data into the deep learning model for health risk assessment processing to obtain a predicted second health risk assessment result; determining a loss value based on the first health risk assessment result in the current sample data and the predicted second health risk assessment result; when the loss value is greater than a preset second loss value, adjusting the parameters of the deep learning model to update the deep learning model, and returning to execute the step of selecting a second sample data from multiple second sample data as the current sample data; when the loss value is less than or equal to the preset second loss value, stopping training the deep learning model to obtain a health risk assessment model.

[0058] Step S105: Verify the initial underwriting result based on the health risk assessment result and the target physiological data, and determine the target underwriting result for the target customer based on the verification result of the initial underwriting result.

[0059] This embodiment uses a pre-trained machine learning model as the target underwriting model to automatically determine the initial underwriting result of the target customer based on the underwriting-related data of the target customer. Then, when the initial underwriting result is passed, the pre-trained deep learning model is used as the health risk assessment model to automatically determine the health risk assessment result of the target customer based on the medical image data of the target customer. Based on the health risk assessment result and the target physiological data, the initial underwriting result is further verified. Finally, according to the verification result, the target underwriting result of the target customer is determined. The entire underwriting process does not require the participation of insurance company staff, reduces the impact of human factors on underwriting, improves the efficiency and accuracy of underwriting, and the underwriting result is further verified based on the health risk assessment result and the target physiological data, further improving the accuracy of underwriting, thereby reducing the risk of insurance company claims.

[0060] In some embodiments, verifying the initial underwriting result based on the health risk assessment result and the target physiological data may include: determining a health risk score of the target customer based on the health risk assessment result and the target physiological data; determining that the initial underwriting result has passed verification in response to the health risk score being less than or equal to a preset score threshold; and determining that the initial underwriting result has failed verification in response to the health risk score being greater than the preset score threshold. The target underwriting result includes either underwriting passed or underwriting failed. If the verification result of the initial underwriting result is that the initial underwriting result has passed verification, the target underwriting result is underwriting passed; if the verification result of the initial underwriting result is that the initial underwriting result has failed verification, the target underwriting result is underwriting failed.

[0061] In some embodiments, determining a health risk score for a target customer based on the health risk assessment results and the target physiological data may include: determining the health risk score for the target customer using a preset health risk scoring model based on the health risk assessment results and the target physiological data. A higher health risk score indicates a higher health risk for the target customer, and a lower health risk score indicates a lower health risk for the target customer. The health risk scoring model is obtained by pre-training a neural network model based on a plurality of third sample data, where the third sample data includes a health risk assessment result sample, a target physiological data sample, and a labeled first health risk score.

[0062] In some embodiments, training a neural network model based on multiple third sample data to obtain a health risk scoring model may include: initializing the neural network model; selecting a third sample data from multiple third sample data as current sample data; inputting the health risk assessment result sample and the target physiological data sample in the current sample data into the neural network model for health risk scoring processing to obtain a predicted second health risk score; determining a loss value based on the first health risk score in the current sample data and the predicted second health risk score; when the loss value is greater than a preset third loss value, adjusting the parameters of the neural network model to update the neural network model, and returning to execute the step of selecting a third sample data from multiple third sample data as current sample data; when the loss value is less than or equal to the preset third loss value, stopping training the neural network model to obtain a health risk scoring model.

[0063] See also Figure 4 , Figure 4 This is a flowchart of the steps of another artificial intelligence-based underwriting method provided in an embodiment of the present application.

[0064] like Figure 4 As shown, the underwriting method includes steps S201 to S206.

[0065] Step S201: Obtain underwriting-related data of the target customer to be underwritten.

[0066] In this embodiment, the target customer may be a customer who newly purchases life insurance products such as a new contract, and the underwriting-related data is data related to underwriting, which may include the target customer's age, gender, occupation, health behavior data obtained by manual inquiry of the target customer, financial information, health status information, and insurance history information, etc.

[0067] Step S202: Determine the initial underwriting result of the target customer based on underwriting-related data using a preset target underwriting model, where the target underwriting model is a pre-trained machine learning model.

[0068] In this embodiment, the pre-trained machine learning model may include a pre-trained neural network model or a gradient boosting tree model. For example, the target underwriting model is obtained by pre-training a neural network model based on a plurality of first sample data, where the first sample data includes underwriting-related data samples and annotated first underwriting results.

[0069] Step S203: In response to the initial underwriting result being approved, obtain the target customer's medical image data, target physiological data, and health behavior data.

[0070] In this embodiment, medical image data includes X-rays, computed tomography (CT) images, single-photon emission computed tomography (SPECT) images, magnetic resonance imaging (MRI), functional magnetic resonance imaging (fMRI), positron emission tomography (PET) images, ultrasound images, and microscopic images. The difference between the acquisition time of the target physiological data and the current time is less than or equal to a preset difference, and the target physiological data includes heart rate, blood pressure, blood oxygen, etc. Health behavior data includes exercise habits, sleep habits, eating habits, drinking habits, smoking habits, etc.

[0071] In some embodiments, in response to the initial underwriting result being an underwriting pass, obtaining the target customer's medical image data, target physiological data, and health behavior data may include: in response to the initial underwriting result being an underwriting pass, obtaining the target customer's medical image data through a second server; obtaining the target physiological data from a physiological database of the target customer stored on the second server, wherein the difference between the acquisition time of the target physiological data and the current time is less than or equal to a preset difference, and the physiological data in the physiological database is collected through a wearable device worn by the target customer; and obtaining the target customer's health behavior data from a third server, the third server including an intelligent consultation module, and the health behavior data is collected through the intelligent consultation module.

[0072] In some embodiments, the intelligent consultation module is deployed with a BERT model that has been pre-trained based on multiple question and answer sample data in the medical and health field. The question and answer sample data includes medical and health questions and labeled answers to the questions. Among them, the third server also stores a medical knowledge graph. When the third server receives the medical and health questions sent by the terminal device, it outputs the corresponding answer text for the medical and health questions through the BERT model combined with the medical knowledge graph, and sends the answer text to the terminal device so that the terminal device displays the answer text in the intelligent consultation dialog box. In this embodiment, customers can consult the third server about medical and health issues through the terminal device, and the third server provides personalized health advice and disease prevention plans, which provides a better customer experience.

[0073] Step S204: Determine the health risk assessment result of the target customer based on the medical image data through a preset health risk assessment model, where the health risk assessment model is a pre-trained deep learning model.

[0074] In this embodiment, the pre-trained deep learning models may include convolutional neural network models (CNN), transformer models, recurrent neural network models (RNN), and deep convolutional neural network models (DNN).

[0075] Step S205: Verify the initial underwriting result based on the health risk assessment result, target physiological data and health behavior data, and determine the target underwriting result of the target customer based on the verification result of the initial underwriting result.

[0076] This embodiment uses a pre-trained machine learning model as the target underwriting model to automatically determine the initial underwriting result of the target customer based on the underwriting-related data of the target customer. Then, when the initial underwriting result is passed, the pre-trained deep learning model is used as the health risk assessment model to automatically determine the health risk assessment result of the target customer based on the medical image data of the target customer. The initial underwriting result is further verified based on the health risk assessment result, target physiological data and health behavior data. Finally, based on the verification result, the target underwriting result of the target customer is determined. The entire underwriting process does not require the participation of insurance company staff, reduces the impact of human factors on underwriting, improves the efficiency and accuracy of underwriting, and the underwriting result is further verified based on the health risk assessment result, target physiological data and health behavior data, further improving the accuracy of underwriting, thereby reducing the risk of insurance company claims.

[0077] In some embodiments, verifying the initial underwriting result based on the health risk assessment results, target physiological data, and health behavior data may include: determining a health risk score for the target customer based on the health risk assessment results, target physiological data, and health behavior data; determining that the initial underwriting result has passed verification in response to the health risk score being less than or equal to a preset score threshold; and determining that the initial underwriting result has failed verification in response to the health risk score being greater than the preset score threshold. The preset score threshold may be set based on actual circumstances and is not specifically limited in this embodiment of the present application.

[0078] In some embodiments, determining the health risk score of the target customer based on the health risk assessment results, the target physiological data, and the health behavior data may include: determining the health risk score of the target customer based on the health risk assessment results, the target physiological data, and the health behavior data using a preset health risk scoring model. A higher health risk score indicates a higher health risk for the target customer, and a lower health risk score indicates a lower health risk for the target customer. The health risk scoring model is obtained by pre-training a neural network model based on a plurality of third sample data, where the third sample data includes a health risk assessment result sample, a target physiological data sample, a health behavior data sample, and a labeled first health risk score.

[0079] In some embodiments, training a neural network model based on multiple third sample data to obtain a health risk scoring model may include: initializing the neural network model; selecting a third sample data from multiple third sample data as current sample data; inputting the health risk assessment result sample, target physiological data sample, and health behavior data sample in the current sample data into the neural network model for health risk scoring processing to obtain a predicted second health risk score; determining a loss value based on the first health risk score in the current sample data and the predicted second health risk score; when the loss value is greater than a preset third loss value, adjusting the parameters of the neural network model to update the neural network model, and returning to execute the step of selecting a third sample data from multiple third sample data as current sample data; when the loss value is less than or equal to the preset third loss value, stopping training the neural network model to obtain a health risk scoring model.

[0080] In some embodiments, before step S101 or step S201, it also includes: obtaining the insurance needs, medical image data, target physiological data and health behavior data of the potential insured customer; determining the health risk assessment result of the potential insured customer based on the medical image data through a preset health risk assessment model, where the health risk assessment model is a pre-trained deep learning model; determining the health risk score of the target customer based on the health risk assessment result, target physiological data and health behavior data; in response to the health risk score being less than or equal to the preset risk score, screening out insurance products that meet the insurance needs from the preset insurance product library as target insurance products and recommending them to the potential insured customer. This embodiment determines the health risk score of the target customer based on the health risk assessment result, target physiological data and health behavior data, and recommends suitable insurance products to the potential insured customer based on the health risk score, thereby improving the user experience and also increasing the success rate of insurance.

[0081] In some embodiments, a potential insured customer can be any customer. Alternatively, a potential insured customer refers to a customer whose number of target policies is less than or equal to a preset number, where the target policies are valid policies stored in the policy database. Alternatively, a potential insured customer includes a customer who has browsed an insurance product but has not purchased the product. The preset number can be set based on actual circumstances and is not specifically limited in this embodiment of the present application. For example, the preset number can be 0, 1, or 2.

[0082] See also Figure 5 , Figure 5 This is a flowchart of the steps of another artificial intelligence-based underwriting method provided in an embodiment of the present application.

[0083] like Figure 5 As shown, after step S105 or step S205, the underwriting method further includes steps S301 to S304.

[0084] Step S301: In response to the target underwriting result being approved, obtain premium pricing related data for the target customer.

[0085] In this embodiment, premium pricing-related data refers to data related to premium pricing, and may include the target customer's age, gender, occupation, financial information, health behavior data, and health status information. Premium pricing-related data may also include insurance type, coverage period, insured amount, payment method, additional liabilities, the insurance company's predetermined interest rate and expense rate, etc. In response to the target underwriting result being a failure, a notification message indicating the failure is pushed to the target customer.

[0086] Step S302: Determine the initial premium based on premium pricing related data using a preset premium pricing model, where the premium pricing model is a pre-trained machine learning model.

[0087] In this embodiment, the pre-trained machine learning model may include a pre-trained neural network model or a gradient boosting tree model. For example, the premium pricing model is obtained by pre-training a neural network model based on a plurality of fourth sample data, where the fourth sample data includes premium pricing-related data samples and the labeled first premium.

[0088] Specifically, the training process of the above-mentioned neural network model may include: initializing the neural network model; selecting a fourth sample data from multiple fourth sample data as the current sample data; inputting the premium pricing-related data sample in the current sample data into the neural network model for premium pricing processing to obtain a predicted second premium; determining the loss value based on the first premium in the current sample data and the predicted second premium; when the loss value is greater than a preset fourth loss value, adjusting the parameters of the neural network model to update the neural network model, and returning to execute the step of selecting a fourth sample data from multiple fourth sample data as the current sample data; when the loss value is less than or equal to the preset fourth loss value, stopping training the neural network model to obtain a premium pricing model.

[0089] Step S303: Determine the health risk score of the target customer based on the health risk assessment result and the target physiological data.

[0090] In this embodiment, a health risk score for a target customer can be determined based on the health risk assessment results and the target physiological data using a preset health risk scoring model. A higher health risk score indicates a higher health risk for the target customer, while a lower health risk score indicates a lower health risk for the target customer. The health risk scoring model is obtained by pre-training a neural network model based on multiple third sample data, where the third sample data includes a health risk assessment result sample, a target physiological data sample, and a labeled first health risk score.

[0091] In some embodiments, determining the health risk score of the target customer based on the health risk assessment results and the target physiological data may include: obtaining the health behavior data of the target customer from a third server; and determining the health risk score of the target customer based on the health risk assessment results, the target physiological data, and the health behavior data. Among them, the health risk score of the target customer can be determined based on the health risk assessment results, the target physiological data, and the health behavior data through a preset health risk scoring model. The health risk scoring model is obtained by pre-training a neural network model based on a plurality of third sample data, and the third sample data includes a health risk assessment result sample, a target physiological data sample, a health behavior data sample, and a labeled first health risk score. This embodiment can more accurately determine the health risk score of the target customer through the health risk assessment results, the target physiological data, and the health behavior data, thereby improving the accuracy of the health risk score.

[0092] Step S304: Determine the premium adjustment coefficient based on the health risk score, and determine the target premium based on the premium adjustment coefficient and the initial premium.

[0093] This embodiment dynamically generates premium adjustment coefficients based on health risk scores and combines them with the initial premium output by the machine learning model to achieve personalized premium pricing, improve customer fairness and retention rates. Through dynamic adjustments, insurance companies can proactively manage high-risk customers and reduce overall claims ratios, while providing discounts to low-risk customers, optimizing business portfolios, and increasing underwriting profit margins. Furthermore, the underwriting method provided by this embodiment forms a closed loop (initial premium → risk score → adjustment coefficient → target premium), reducing the need for manual review, lowering operating costs, and improving the efficiency, flexibility, and accuracy of premium pricing.

[0094] In some embodiments, determining a premium adjustment coefficient based on the health risk score, and determining a target premium based on the premium adjustment coefficient and the initial premium, may include: determining a preset score range for the health risk score, obtaining a preset adjustment coefficient corresponding to the preset score range, and obtaining the premium adjustment coefficient; and multiplying the initial premium by the premium adjustment coefficient to obtain the target premium. Different preset score ranges correspond to different preset adjustment coefficients, and the premium adjustment coefficient can be greater than 1, equal to 1, or even greater than 1.

[0095] See also Figure 6 , Figure 6 It is a schematic block diagram of an artificial intelligence-based underwriting device provided in an embodiment of the present application.

[0096] like Figure 6 As shown, the artificial intelligence-based underwriting device 100 includes:

[0097] An acquisition module 110 is used to acquire underwriting-related data of a target customer to be underwritten;

[0098] An underwriting module 120 is configured to determine an initial underwriting result for the target customer based on the underwriting-related data using a preset target underwriting model, wherein the target underwriting model is a pre-trained machine learning model;

[0099] The acquisition module 110 is further configured to acquire the medical image data and target physiological data of the target customer in response to the initial underwriting result being approved;

[0100] A health risk assessment module 130 is configured to determine a health risk assessment result of the target customer based on the medical image data using a preset health risk assessment model, wherein the health risk assessment model is a pre-trained deep learning model;

[0101] a verification module 140, configured to verify the initial underwriting result based on the health risk assessment result and the target physiological data;

[0102] The determination module 150 is used to determine the target underwriting result of the target customer based on the verification result of the initial underwriting result.

[0103] In some embodiments, the acquisition module 110 is further configured to acquire the medical image data, target physiological data, and health behavior data of the target customer in response to the initial underwriting result being approved;

[0104] The verification module 140 is further configured to verify the initial underwriting result based on the health risk assessment result, the target physiological data, and the health behavior data.

[0105] In some embodiments, the acquisition module 110 is further configured to:

[0106] In response to the initial underwriting result being an approved underwriting, obtaining medical image data of the target customer through a second server;

[0107] acquiring target physiological data from a physiological database of the target customer stored in the second server, wherein a difference between a collection time of the target physiological data and a current time is less than or equal to a preset difference, and the physiological data in the physiological database is collected through a wearable device worn by the target customer; and

[0108] The health behavior data of the target customer is obtained from a third server, where the third server includes an intelligent diagnosis module, and the health behavior data is collected by the intelligent diagnosis module.

[0109] In some embodiments, the verification module 140 is further configured to:

[0110] Determining a health risk score for the target customer based on the health risk assessment result, the target physiological data, and the health behavior data;

[0111] In response to the health risk score being less than or equal to a preset score threshold, determining that the initial underwriting result passes verification;

[0112] In response to the health risk score being greater than the preset score threshold, it is determined that the initial underwriting result has failed verification.

[0113] In some embodiments, the artificial intelligence-based underwriting device 100 further includes:

[0114] The acquisition module 110 is further configured to acquire premium pricing related data of the target customer in response to the target underwriting result being underwriting approval;

[0115] a premium pricing module, configured to determine an initial premium based on the premium pricing-related data using a preset premium pricing model, wherein the premium pricing model is a pre-trained machine learning model;

[0116] The determination module 150 is further configured to determine the health risk score of the target customer based on the health risk assessment result and the target physiological data;

[0117] The determination module 150 is further configured to determine a premium adjustment coefficient based on the health risk score, and determine a target premium based on the premium adjustment coefficient and the initial premium.

[0118] In some embodiments, the determining module 150 is further configured to:

[0119] Obtaining the health behavior data of the target customer from a third server;

[0120] A health risk score of the target customer is determined based on the health risk assessment result, the target physiological data, and the health behavior data.

[0121] In some embodiments, the artificial intelligence-based underwriting device 100 further includes:

[0122] The acquisition module 110 is further used to obtain the insurance needs, medical image data, target physiological data and health behavior data of potential insured customers;

[0123] The health risk assessment module 130 is further configured to determine a health risk assessment result of the potential insured customer based on the medical image data using a preset health risk assessment model, wherein the health risk assessment model is a pre-trained deep learning model;

[0124] The determination module 150 is further configured to determine the health risk score of the target customer based on the health risk assessment result, the target physiological data, and the health behavior data;

[0125] The product recommendation module is used to, in response to the health risk score being less than or equal to the preset risk score, screen out an insurance product that meets the insurance needs from a preset insurance product library and recommend it to the potential insured customer as a target insurance product.

[0126] It should be noted that those skilled in the art will clearly understand that, for the sake of convenience and brevity of description, the specific working processes of the above-described devices and modules and units can refer to the corresponding processes in the aforementioned embodiment of the artificial intelligence-based underwriting method, and will not be repeated here.

[0127] The apparatus provided in the above embodiment can be implemented in the form of a computer program. The computer program can be used in Figure 7 Run on the server shown.

[0128] See also Figure 7 , Figure 7 This is a schematic block diagram of the structure of a server provided in an embodiment of the present application.

[0129] like Figure 7 As shown, the server includes a processor, a memory, and a network interface connected via a system bus, wherein the memory may include a storage medium and an internal memory.

[0130] The storage medium may store an operating system and a computer program. The computer program includes program instructions that, when executed, enable the processor to perform any of the underwriting methods.

[0131] The processor is used to provide computing and control capabilities to support the operation of the entire server.

[0132] The network interface is used for network communication, such as sending assigned tasks, etc. Those skilled in the art will understand that Figure 7 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the server to which the solution of the present application is applied. The specific server may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0133] It should be understood that the processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc.

[0134] In one embodiment, the processor is configured to execute a computer program stored in the memory to implement the following steps:

[0135] Obtain underwriting-related data of target customers to be underwritten;

[0136] determining an initial underwriting result for the target customer based on the underwriting-related data using a preset target underwriting model, wherein the target underwriting model is a pre-trained machine learning model;

[0137] In response to the initial underwriting result being an approved underwriting, obtaining medical image data and target physiological data of the target customer;

[0138] Determining a health risk assessment result of the target customer based on the medical image data using a preset health risk assessment model, wherein the health risk assessment model is a pre-trained deep learning model;

[0139] The initial underwriting result is verified based on the health risk assessment result and the target physiological data, and the target underwriting result of the target customer is determined based on the verification result of the initial underwriting result.

[0140] In some embodiments, when the processor acquires the medical image data and target physiological data of the target customer in response to the initial underwriting result being an approved underwriting, it is configured to:

[0141] In response to the initial underwriting result being an approved underwriting, obtaining medical image data, target physiological data, and health behavior data of the target customer;

[0142] Verifying the initial underwriting result based on the health risk assessment result and the target physiological data includes:

[0143] The initial underwriting result is verified based on the health risk assessment result, the target physiological data and the health behavior data.

[0144] In some embodiments, when the processor acquires the medical image data, target physiological data, and health behavior data of the target customer in response to the initial underwriting result being an approved underwriting, it is configured to:

[0145] In response to the initial underwriting result being an approved underwriting, obtaining medical image data of the target customer through a second server;

[0146] acquiring target physiological data from a physiological database of the target customer stored in the second server, wherein a difference between a collection time of the target physiological data and a current time is less than or equal to a preset difference, and the physiological data in the physiological database is collected through a wearable device worn by the target customer; and

[0147] The health behavior data of the target customer is obtained from a third server, where the third server includes an intelligent diagnosis module, and the health behavior data is collected by the intelligent diagnosis module.

[0148] In some embodiments, when verifying the initial underwriting result based on the health risk assessment result, the target physiological data, and the health behavior data, the processor is configured to implement:

[0149] Determining a health risk score for the target customer based on the health risk assessment result, the target physiological data, and the health behavior data;

[0150] In response to the health risk score being less than or equal to a preset score threshold, determining that the initial underwriting result passes verification;

[0151] In response to the health risk score being greater than the preset score threshold, it is determined that the initial underwriting result has failed verification.

[0152] In some embodiments, after determining the target underwriting result for the target customer based on the verification result of the initial underwriting result, the processor is further configured to:

[0153] In response to the target underwriting result being an approved underwriting, obtaining premium pricing related data for the target customer;

[0154] Determining an initial premium based on the premium pricing related data using a preset premium pricing model, wherein the premium pricing model is a pre-trained machine learning model;

[0155] Determining a health risk score for the target customer based on the health risk assessment result and the target physiological data;

[0156] A premium adjustment coefficient is determined based on the health risk score, and a target premium is determined based on the premium adjustment coefficient and the initial premium.

[0157] In some embodiments, when determining the health risk score of the target customer based on the health risk assessment result and the target physiological data, the processor is configured to implement:

[0158] Obtaining the health behavior data of the target customer from a third server;

[0159] A health risk score of the target customer is determined based on the health risk assessment result, the target physiological data, and the health behavior data.

[0160] In some embodiments, before obtaining the underwriting-related data of the target customer to be underwritten, the processor is further configured to:

[0161] Obtain insurance needs, medical image data, target physiological data, and health behavior data of potential insured customers;

[0162] Determining a health risk assessment result of the potential insured customer based on the medical image data using a preset health risk assessment model, wherein the health risk assessment model is a pre-trained deep learning model;

[0163] Determining a health risk score for the target customer based on the health risk assessment result, the target physiological data, and the health behavior data;

[0164] In response to the health risk score being less than or equal to the preset risk score, an insurance product that meets the insurance needs is screened out from a preset insurance product library and recommended to the potential insured customer as a target insurance product.

[0165] It should be noted that those skilled in the art will clearly understand that, for the sake of convenience and brevity of description, the specific working process of the server described above can refer to the corresponding process in the aforementioned embodiment of the artificial intelligence-based underwriting method, and will not be repeated here.

[0166] Through the description of the above implementation methods, it can be seen that those skilled in the art can clearly understand that the present application can be implemented by means of software plus a necessary general hardware platform. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product can be stored in a storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a server (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present application or certain parts of the embodiments.

[0167] An embodiment of the present application also provides a storage medium, on which a computer program is stored. The computer program includes program instructions. The method implemented when the program instructions are executed can be referred to the various embodiments of the underwriting method of the present application.

[0168] The storage medium may be volatile or non-volatile. The storage medium may be an internal storage unit of the server described in the aforementioned embodiment, such as a hard disk or memory of the server. The storage medium may also be an external storage device of the server, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the server.

[0169] Furthermore, the storage medium may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function, etc.; the data storage area may store data created according to the use of the blockchain node, etc.

[0170] The blockchain referred to in this application is a new application model for computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanisms, and encryption algorithms. Blockchain is essentially a decentralized database, a series of data blocks generated using cryptographic methods. Each data block contains information about a batch of network transactions, which is used to verify the validity of this information (to prevent counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, the platform product service layer, and the application service layer.

[0171] It should be noted that in each specific implementation method of this application, when it comes to the need to process the customer's underwriting-related data, premium pricing-related data, medical image data, physiological data, health behavior data and other data related to the customer's identity or characteristics, the user's permission or consent will be obtained first, and the collection, use and processing of these data will comply with relevant laws, regulations and standards.

[0172] It should be understood that any non-Company software tools or components that appear in the embodiments of this application are merely examples and do not represent actual use. The terms used in this application specification are merely for the purpose of describing specific embodiments and are not intended to limit this application. As used in this application specification and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms unless the context clearly indicates otherwise.

[0173] It should also be understood that the term "and / or" used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, including these combinations. It should be noted that, in this article, the terms "include", "comprise" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system that includes a series of elements includes not only those elements, but also other elements that are not explicitly listed, or also includes elements that are inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "including a..." does not exclude the presence of other identical elements in the process, method, article or system that includes the element.

[0174] The serial numbers of the embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments. The above description is only a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in this application, and these modifications or replacements should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. An artificial intelligence-based underwriting method, characterized in that: Applied to the first server, the method includes: Obtain underwriting-related data of target customers to be underwritten; determining an initial underwriting result for the target customer based on the underwriting-related data using a preset target underwriting model, wherein the target underwriting model is a pre-trained machine learning model; In response to the initial underwriting result being an approved underwriting, obtaining medical image data and target physiological data of the target customer; Determining a health risk assessment result of the target customer based on the medical image data using a preset health risk assessment model, wherein the health risk assessment model is a pre-trained deep learning model; The initial underwriting result is verified based on the health risk assessment result and the target physiological data, and the target underwriting result of the target customer is determined based on the verification result of the initial underwriting result.

2. The underwriting method according to claim 1, characterized in that: In response to the initial underwriting result being an approved underwriting, obtaining medical image data and target physiological data of the target customer includes: In response to the initial underwriting result being an approved underwriting, obtaining medical image data, target physiological data, and health behavior data of the target customer; Verifying the initial underwriting result based on the health risk assessment result and the target physiological data includes: The initial underwriting result is verified based on the health risk assessment result, the target physiological data and the health behavior data.

3. The underwriting method according to claim 2, characterized in that: In response to the initial underwriting result being an approved underwriting, obtaining the medical image data, target physiological data, and health behavior data of the target customer includes: In response to the initial underwriting result being an approved underwriting, obtaining medical image data of the target customer through a second server; acquiring target physiological data from a physiological database of the target customer stored in the second server, wherein a difference between a collection time of the target physiological data and a current time is less than or equal to a preset difference, and the physiological data in the physiological database is collected through a wearable device worn by the target customer; and The health behavior data of the target customer is obtained from a third server, where the third server includes an intelligent diagnosis module, and the health behavior data is collected by the intelligent diagnosis module.

4. The underwriting method according to claim 2, characterized in that: The verifying of the initial underwriting result based on the health risk assessment result, the target physiological data, and the health behavior data includes: Determining a health risk score for the target customer based on the health risk assessment result, the target physiological data, and the health behavior data; In response to the health risk score being less than or equal to a preset score threshold, determining that the initial underwriting result passes verification; In response to the health risk score being greater than the preset score threshold, it is determined that the initial underwriting result has failed verification.

5. The underwriting method according to any one of claims 1 to 4, characterized in that: After determining the target underwriting result for the target customer based on the verification result of the initial underwriting result, the method further includes: In response to the target underwriting result being an approved underwriting, obtaining premium pricing related data for the target customer; Determining an initial premium based on the premium pricing related data using a preset premium pricing model, wherein the premium pricing model is a pre-trained machine learning model; Determining a health risk score for the target customer based on the health risk assessment result and the target physiological data; A premium adjustment coefficient is determined based on the health risk score, and a target premium is determined based on the premium adjustment coefficient and the initial premium.

6. The underwriting method according to claim 5, characterized in that: Determining the health risk score of the target customer based on the health risk assessment result and the target physiological data includes: Obtaining the health behavior data of the target customer from a third server; A health risk score of the target customer is determined based on the health risk assessment result, the target physiological data, and the health behavior data.

7. The underwriting method according to any one of claims 1 to 4, characterized in that: Before obtaining the underwriting-related data of the target customer to be underwritten, the following steps are also included: Obtain insurance needs, medical image data, target physiological data, and health behavior data of potential insured customers; Determining a health risk assessment result of the potential insured customer based on the medical image data using a preset health risk assessment model, wherein the health risk assessment model is a pre-trained deep learning model; Determining a health risk score for the target customer based on the health risk assessment result, the target physiological data, and the health behavior data; In response to the health risk score being less than or equal to the preset risk score, an insurance product that meets the insurance needs is screened out from a preset insurance product library and recommended to the potential insured customer as a target insurance product.

8. An artificial intelligence-based underwriting device, characterized in that: The underwriting device includes: The acquisition module is used to obtain underwriting-related data of the target customers to be underwritten; an underwriting module, configured to determine an initial underwriting result for the target customer based on the underwriting-related data using a preset target underwriting model, wherein the target underwriting model is a pre-trained machine learning model; The acquisition module is further configured to acquire the medical image data and target physiological data of the target customer in response to the initial underwriting result being an approved underwriting; A health risk assessment module, configured to determine a health risk assessment result of the target customer based on the medical image data using a preset health risk assessment model, wherein the health risk assessment model is a pre-trained deep learning model; a verification module, configured to verify the initial underwriting result based on the health risk assessment result and the target physiological data; A determination module is used to determine the target underwriting result of the target customer based on the verification result of the initial underwriting result.

9. A server, characterized in that: The server includes a processor, a memory, and a computer program stored in the memory and executable by the processor, wherein when the computer program is executed by the processor, the steps of the artificial intelligence-based underwriting method as described in any one of claims 1 to 7 are implemented.

10. A storage medium for computer-readable storage, characterized in that: The storage medium stores a computer program, wherein when the computer program is executed by the processor, the steps of the artificial intelligence-based underwriting method as described in any one of claims 1 to 7 are implemented.

Images