A kind of underwriting risk assessment method, device, equipment and medium
Patent Information
- Application Number
- CN202610663038.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-05-13
- Publication Date
- 2026-09-22
AI Technical Summary
[0003]鉴于此，本申请实施例提供了一种核保风险评估方法、装置、设备及介质，以解决在核保过程中，对投保人核保风险评估的准确性低的问题
本申请中，基于第一神经网络模型对投保数据进行处理，确定投保数据的多个特征与每个特征对应的特征权重值；确定每个特征的重要性评分值，根据每个特征的特征权重值和对应的重要性评分值，计算得到投保人的第一核保风险概率值，基于第二神经网络模型对投保数据进行处理，确定投保人的第二核保风险概率值；根据第一核保风险概率值与第二核保风险概率值，对投保人的核保风险进行等级分类，确定投保人的核保风险结果，融合第一核保风险概率值与第二核保风险概率值，进行核保风险等级分类，考虑了预测过程中特征重要程度不同情况下的预测结果，从而提高了最终的核保风险概率值的准确性。
Smart Images
Figure CN122798546A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of artificial intelligence technology, and in particular to an underwriting risk assessment method, apparatus, equipment and medium. Background Technology
[0002] Underwriting is the process by which insurance companies assess and analyze an applicant's health status, financial situation, occupational risks, and other information to decide whether to underwrite the policy and determine the terms and premiums. Underwriting is an indispensable part of the insurance industry; it not only protects the interests of insurance companies but also provides policyholders with professional assessment and protection. In the insurance industry, underwriting is a crucial step in determining whether a policy is accepted and how the premium is determined. Currently, underwriting is conducted manually, lacking dynamic analysis of multi-dimensional data on the applicant's health, finances, and behavior. This results in low accuracy in assessing underwriting risk, leading to a high rate of missed applications from high-risk clients. Therefore, improving the accuracy of underwriting risk assessment is a pressing issue that needs to be addressed. Summary of the Invention
[0003] In view of this, embodiments of this application provide an underwriting risk assessment method, apparatus, equipment, and medium to solve the problem of low accuracy in assessing the underwriting risk of the insured during the underwriting process.
[0004] In a first aspect, embodiments of this application provide an underwriting risk assessment method, the underwriting risk assessment method comprising: Obtain the policyholder's insurance application data, process the insurance application data based on the first neural network model, and determine multiple features of the insurance application data and the feature weight value corresponding to each feature; Determine the importance score for each feature, and calculate the first underwriting risk probability value for the policyholder based on the feature weight value and the corresponding importance score for each feature. The insurance data is processed based on the second neural network model to determine the second underwriting risk probability value of the policyholder; Based on the first underwriting risk probability value and the second underwriting risk probability value, the underwriting risk of the policyholder is classified into different levels, and the underwriting risk result of the policyholder is determined.
[0005] Secondly, embodiments of this application provide an underwriting risk assessment device, the underwriting risk assessment device comprising: The first determining module is used to acquire the policyholder's insurance data, process the insurance data based on the first neural network model, and determine multiple features of the insurance data and the feature weight value corresponding to each feature; The calculation module is used to determine the importance score of each feature, and calculate the first underwriting risk probability value of the policyholder based on the feature weight value and the corresponding importance score of each feature. The second determining module is used to process the insurance application data based on the second neural network model to determine the second underwriting risk probability value of the policyholder; The classification module is used to classify the underwriting risk of the policyholder according to the first underwriting risk probability value and the second underwriting risk probability value, and to determine the underwriting risk result of the policyholder.
[0006] Thirdly, embodiments of this application provide a computer device, the computer device including a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the underwriting risk assessment method as described above.
[0007] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the underwriting risk assessment method as described above.
[0008] The advantages of this application compared to the prior art are: In this application, the insurance application data is processed based on a first neural network model to determine multiple features of the data and the corresponding feature weights for each feature; the importance score for each feature is determined; and based on the feature weights and corresponding importance scores, the first underwriting risk probability value for the insured is calculated. The insurance application data is then processed based on a second neural network model to determine the second underwriting risk probability value for the insured. Based on the first and second underwriting risk probability values, the insured's underwriting risk is classified into levels to determine the underwriting risk result. Finally, the first and second underwriting risk probability values are merged to classify the underwriting risk level, taking into account the prediction results under different feature importance levels during the prediction process, thereby improving the accuracy of the final underwriting risk probability value. Attached Figure Description
[0009] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0010] Figure 1 This is a schematic diagram illustrating the application environment of an underwriting risk assessment method provided in an embodiment of this application; Figure 2 This is a flowchart illustrating an underwriting risk assessment method provided in one embodiment of this application; Figure 3 This is a schematic diagram of the structure of an underwriting risk assessment device provided in one embodiment of this application; Figure 4 This is a schematic diagram of the structure of a computer device provided in one embodiment of this application. Detailed Implementation
[0011] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0012] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.
[0013] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.
[0014] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.
[0015] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if detected [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once detected [the described condition or event]," or "in response to detection [the described condition or event]."
[0016] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.
[0017] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.
[0018] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, technologies, 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 obtain optimal results.
[0019] Foundational technologies for artificial intelligence generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies mainly encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.
[0020] It should be understood that the sequence number of each step in the following embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.
[0021] To illustrate the technical solution of this application, specific embodiments are described below.
[0022] One embodiment of this application provides an underwriting risk assessment method that can be applied to, for example... Figure 1In this application environment, the client communicates with the server. The client includes, but is not limited to, handheld computers, desktop computers, laptops, ultra-mobile personal computers (UMPCs), netbooks, cloud computing devices, and personal digital assistants (PDAs). The server can be a standalone server or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.
[0023] To illustrate the technical solution of this application, specific embodiments are described below.
[0024] See Figure 2 This is a flowchart illustrating an underwriting risk assessment method provided in an embodiment of this application, as shown below. Figure 2 As shown, the underwriting risk assessment method may include the following steps.
[0025] S201: Obtain the policyholder's insurance data, process the insurance data based on the first neural network model, and determine multiple features of the insurance data and the feature weight value corresponding to each feature.
[0026] In step S201, the policyholder's insurance data includes the policyholder's multi-dimensional data. The first neural network model is used to determine multiple features of the insurance data and the feature weight value corresponding to each feature. The multiple features are the features of the policyholder's multi-dimensional data, and the feature weight value is the coefficient of the model parameter directly obtained after processing the insurance data based on the first neural network model.
[0027] In this embodiment, the policyholder's insurance application data is multi-dimensional data, which may include health-related data (such as medical examination reports and medical records), financial-related data (such as income level and debt situation), behavioral-related data (such as insurance application frequency and historical claims records), and external-related data (such as credit records and social media behavior). The first neural network model is an XGBoost (eXtreme GradientBoosting) model. The XGBoost model is used to process the insurance application data to determine multiple features of the data. Each feature and its corresponding feature weight value represents a feature of a dimension, and the feature weight value is the weight value obtained by processing the insurance application data using the XGBoost model, i.e., the feature's contribution to the tree structure. The process of the XGBoost model processing the insurance application data is not limited in this embodiment. The XGBoost model is a pre-trained model; the insurance application data is input into the XGBoost model to obtain multiple features of the insurance application data and their corresponding feature weight values.
[0028] For example, when an insured person applies for an insurance product, in order to underwrite the insured person's application data, data such as the insured person's medical examination report, medical records, income level, debt situation, frequency of insurance applications, historical claims records, credit records and social media behavior are obtained. Based on the first neural network model, the data such as medical examination report, medical records, income level, debt situation, frequency of insurance applications, historical claims records, credit records and social media behavior are processed to determine multiple features and the feature weight value corresponding to each feature.
[0029] In this embodiment, the policyholder's insurance data is obtained, and the insurance data is processed based on the first neural network model to determine multiple features of the insurance data and the feature weight value corresponding to each feature, so as to understand the importance of insurance data in different dimensions and the degree of influence on the target variable.
[0030] S202: Determine the importance score for each feature, and calculate the policyholder's first underwriting risk probability value based on the feature weight value and the corresponding importance score for each feature.
[0031] In step S202, the importance score of each feature is the degree of importance of each feature to the underwriting risk prediction. Based on the feature weight value and the corresponding importance score of each feature, the first underwriting risk probability value of the policyholder is calculated, where the first underwriting risk probability value represents the size of the policyholder's insurance risk.
[0032] In this embodiment, when determining the importance score of each feature, the importance score of each feature is the SHAP (SHapley Additive exPlanations) value of each feature, which characterizes the importance of each feature to underwriting risk prediction. The formula for calculating the first underwriting risk probability value is as follows: in, This is the probability value for the first underwriting risk. The importance score for the i-th feature. Let be the feature weight value of the i-th feature, and m be the number of features.
[0033] In this embodiment, the importance score of each feature is determined, and the first underwriting risk probability value of the policyholder is calculated based on the feature weight value and the corresponding importance score value of each feature. Taking into account the feature weight value and the corresponding importance score value of each feature can significantly improve the interpretability and reliability of the first underwriting risk probability value, thereby improving the accuracy of the first underwriting risk probability value.
[0034] Optionally, an importance score for each feature is determined, including: Obtain training samples, which include insurance application data from multiple policyholders, and determine the sample value corresponding to each feature in each insurance application data. For any feature, select any sample insurance data, and determine the importance score of the feature in the corresponding sample insurance data based on the feature weight value corresponding to each feature and the sample value of each feature in the sample insurance data. Iterate through all sample insurance data to obtain the importance score of the feature in each sample insurance data.
[0035] Based on the importance score of each feature in each sample of insurance data, determine the importance score of the feature; iterate through all features and determine the importance score of each feature.
[0036] In this embodiment, the importance score for each feature is determined using multiple training samples. These training samples include insurance application data from multiple policyholders, and the sample value corresponding to each feature in each insurance application data set is determined. The number of training samples is greater than or equal to the number of features in the insurance application data set.
[0037] For any given feature, select any sample of insurance application data. Based on the feature weight value corresponding to each feature and the value of each feature in the sample insurance application data, determine the importance score of the feature in the corresponding sample insurance application data. The calculation formula is as follows: in, Let m be the feature weight value of the j-th feature, and m be the number of features. For the j-th feature in the sample insurance data The sample values, For feature i in the sample insurance data Importance score in the context.
[0038] Iterate through all insurance application data samples to obtain the importance score of each feature in each sample. Based on the importance score of each feature in each sample, determine the feature's overall importance score. One method to determine the feature's importance score is to sum the importance scores of each feature in each sample and divide by the number of insurance application samples. Repeat this process for all features to determine the importance score for each feature.
[0039] In this embodiment, the importance score of the corresponding feature is calculated based on the insurance data of all samples, taking into account the degree of influence of the feature on the insurance data of each sample, thereby improving the accuracy of calculating the importance score of the feature.
[0040] Optionally, the importance score of the feature is determined based on its importance score in each sample of insurance data, including: Cluster the importance scores of features across all insurance data samples to determine the initial insurance data samples; The difference between the sample value of the feature in each initial sample insurance data and the preset feature threshold is determined. When the difference is greater than the preset difference threshold, the corresponding initial sample insurance data is determined as the target sample insurance data. The mean of the importance scores of the corresponding features in all the insurance data of the target samples is determined as the importance score of the corresponding feature.
[0041] In this embodiment, the importance scores of features across all sample insurance application data are clustered to identify the sample insurance application data corresponding to the most frequent importance scores in each category as the initial sample insurance application data. A preset feature threshold is obtained; different features have different preset feature thresholds, which can be set according to specific circumstances. The difference between the feature's sample value in each initial sample insurance application data and the preset feature threshold is determined. When the difference exceeds the preset difference threshold, the corresponding initial sample insurance application data is identified as the target sample insurance application data. This facilitates the removal of abnormal sample insurance application data and improves the accuracy of calculating the corresponding feature importance score.
[0042] The average importance score of the corresponding feature in all target sample insurance data is determined as the importance score of the corresponding feature. The importance score of the feature in each target sample insurance data is added together and then divided by the number of target sample insurance data to obtain the importance score of the feature.
[0043] In this embodiment, initial sample insurance data is determined by clustering the importance scores of each feature. This identifies similar importance scores for each feature across different sample insurance data, avoiding the inclusion of extreme cases where the feature's importance score might be misrepresented, thus improving the reliability of calculating the feature's importance score from the sample insurance data. Based on the initial sample insurance data, sample insurance data with significant differences between the feature's sample value and its threshold are removed, thereby eliminating abnormal sample insurance data and improving the accuracy of the target sample insurance data. The importance score of the feature is then calculated based on the target sample insurance data, further enhancing the accuracy of the feature's importance score.
[0044] S203: Based on the second neural network model, process the insurance data to determine the second underwriting risk probability value of the policyholder.
[0045] In step S203, the second neural network model is a Transformer or LSTM model, etc. The insurance data is processed based on the second neural network model, that is, the risk prediction of the insurance data is performed based on the second neural network model, and the insurance data is processed based on the second neural network model.
[0046] In this embodiment, the second neural network model is a Transformer model. Based on the Transformer model, complex feature relationships are extracted from the insurance application data, and the weight value corresponding to each feature is determined. Based on the weight value corresponding to each feature, the second underwriting risk probability value for the policyholder is determined. The insurance application data is converted into serialized data and input into the Transformer model, outputting the second underwriting risk probability value. Specifically, after the serialized data is input into the Transformer model, attention weight values of the insurance application data are extracted based on an attention mechanism, and the corresponding second underwriting risk probability value is determined based on these attention weight values.
[0047] In this embodiment, the insurance data is mapped to three subspaces through linear transformation: in, For insurance data, , , is a learnable parameter matrix.
[0048] according to , , The generated attention weight matrix is calculated using the following formula: in, Here is the attention weight matrix. Let X be the dimension.
[0049] The formula for calculating the probability value of the second underwriting risk is as follows: in, This is the second underwriting risk probability value. b represents the model parameters of the Transformer model. This is the attention weight matrix.
[0050] In this embodiment, the corresponding risk probability value is calculated based on the attention mechanism. The corresponding weight value can be dynamically allocated according to the input insurance data, and high-risk characteristics can be automatically identified, thereby improving the accuracy of the second underwriting risk probability value.
[0051] S204: Based on the first underwriting risk probability value and the second underwriting risk probability value, classify the underwriting risk of the policyholder into different levels and determine the underwriting risk result of the policyholder.
[0052] In step S204, the classification is used to categorize the magnitude of underwriting risk and determine the underwriting risk outcome for the policyholder.
[0053] In this embodiment, the underwriting risk of the policyholder is classified into levels according to the first underwriting risk probability and the second underwriting risk probability. When determining the underwriting risk result of the policyholder, the average of the first underwriting risk probability and the second underwriting risk probability can be calculated, and the underwriting risk result of the policyholder can be determined according to the corresponding average.
[0054] When classifying the underwriting risk of the insured, it can be divided into three categories, such as low risk, medium risk and high risk. The final underwriting risk probability value is classified as low risk if it is in the range of 0-0.3, medium risk if it is in the range of 0.3-0.7, and high risk if it is in the range of 0.7-1.
[0055] In this embodiment, the first underwriting risk probability value and the second underwriting risk probability value are integrated to classify the underwriting risk level. The prediction results under different feature importance conditions are taken into account during the prediction process, thereby improving the accuracy of the final underwriting risk probability value.
[0056] Optionally, based on the first underwriting risk probability and the second underwriting risk probability, the insured's underwriting risk is classified into levels to determine the insured's underwriting risk outcome, including: Determine the first weight value of the first underwriting risk probability value and the second weight value of the second underwriting risk probability value; Based on the first weight value and the second weight value, the first underwriting risk probability value and the second underwriting risk probability value are weighted and summed to obtain the target underwriting risk probability value; Based on the target underwriting risk probability value, the underwriting risk of the policyholder is classified into different levels to determine the underwriting risk outcome for the policyholder.
[0057] In this embodiment, different weight values are set for the first underwriting risk probability value and the second underwriting risk probability value. The corresponding weights can be determined based on the accuracy of the first neural network model and the accuracy of the second neural network model. The accuracy of the first neural network model and the accuracy of the second neural network model are determined based on their respective accuracy rates. The corresponding weight values are then calculated using the following formula: in, The accuracy of the first neural network model. The accuracy of the second neural network model. The first weighting value is the first underwriting risk probability value. It is the second weighting value of the second underwriting risk probability value.
[0058] Based on the first weight value and the second weight value, the first underwriting risk probability value and the second underwriting risk probability value are weighted and summed to obtain the target underwriting risk probability value. Based on the target underwriting risk probability value, the underwriting risk of the policyholder is classified into levels to determine the underwriting risk result of the policyholder.
[0059] In this embodiment, based on the corresponding accuracy rate, a first weight value for the first underwriting risk probability value and a second weight value for the second underwriting risk probability value are determined, so that the importance of the first underwriting risk probability value and the second underwriting risk probability value is associated with the accuracy rate of the corresponding neural network model. When the accuracy rate of the neural network model is higher, the corresponding underwriting risk probability value is more accurate. Therefore, based on the corresponding accuracy rate, the first weight value for the first underwriting risk probability value and the second weight value for the second underwriting risk probability value are determined, thereby improving the accuracy of the underwriting risk result.
[0060] Optionally, based on the target underwriting risk probability value, the insured's underwriting risk is classified into levels to determine the insured's underwriting risk outcome, including: Determine the clustering intervals for underwriting risk levels, and establish a correspondence between the underwriting risk level clustering intervals and the underwriting risk results; Based on the clustering intervals of underwriting risk levels, determine the target clustering interval corresponding to the target underwriting risk probability value, and based on the target clustering interval, determine the underwriting risk result corresponding to the target clustering interval.
[0061] In this embodiment, when determining the clustering interval of underwriting risk level, the underwriting probability value of sample data is obtained. The sample data refers to the insurance application data of sample policyholders. The underwriting probability values of the sample data are clustered to determine the clustering interval for each cluster. Based on the clustering interval of underwriting risk level, the target clustering interval corresponding to the target underwriting risk probability value is determined. Based on the target clustering interval, the underwriting risk result corresponding to the target clustering interval is determined.
[0062] In this embodiment, clustering is used to dynamically determine the corresponding underwriting risk level range, thereby improving the accuracy of underwriting risk level classification and thus improving the accuracy of determining the underwriting risk result corresponding to the target clustering range.
[0063] Optionally, after determining the underwriting risk outcome for the policyholder, the following steps are also included: Based on the underwriting risk results, determine the appropriate wording to match the underwriting risk results.
[0064] In this embodiment, the script matched with the underwriting risk result is a customized script used to discourage policyholder attrition and disputes. For example, if the underwriting risk result is high risk, free health management services may be provided, installment payment plans may be recommended, or the sum insured / coverage may be adjusted.
[0065] In this embodiment, based on the underwriting risk results, a script matching the underwriting risk results is determined in order to proactively intervene with high-risk customers and improve the retention rate of high-risk customers.
[0066] In this application, the insurance application data is processed based on a first neural network model to determine multiple features of the data and the corresponding feature weights for each feature; the importance score for each feature is determined; and based on the feature weights and corresponding importance scores, the first underwriting risk probability value for the insured is calculated. The insurance application data is then processed based on a second neural network model to determine the second underwriting risk probability value for the insured. Based on the first and second underwriting risk probability values, the insured's underwriting risk is classified into levels to determine the underwriting risk result. Finally, the first and second underwriting risk probability values are merged to classify the underwriting risk level, taking into account the prediction results under different feature importance levels during the prediction process, thereby improving the accuracy of the final underwriting risk probability value.
[0067] Please see Figure 3 , Figure 3 This is a schematic diagram of an underwriting risk assessment device according to an embodiment of this application. This underwriting risk assessment device corresponds one-to-one with the underwriting risk assessment methods described in the above embodiments. Please refer to [link / reference] for details. Figure 2 as well as Figure 2 The relevant descriptions in the corresponding embodiments are shown below. For ease of explanation, only the parts relevant to this embodiment are shown. See also... Figure 3 The underwriting risk assessment device 30 includes: a first determination module 31, a calculation module 32, a second determination module 33, and a classification module 34.
[0068] The first determining module 31 is used to obtain the policyholder's insurance data, process the insurance data based on the first neural network model, and determine multiple features of the insurance data and the feature weight value corresponding to each feature.
[0069] The calculation module 32 is used to determine the importance score of each feature and calculate the first underwriting risk probability value of the policyholder based on the feature weight value and the corresponding importance score of each feature.
[0070] The second determining module 33 is used to process the insurance data based on the second neural network model to determine the second underwriting risk probability value of the insured.
[0071] The classification module 34 is used to classify the underwriting risk of the policyholder according to the first underwriting risk probability value and the second underwriting risk probability value, and to determine the underwriting risk result of the policyholder.
[0072] Optionally, the computing module includes: The acquisition unit is used to acquire training samples, which include insurance data of multiple policyholders and determine the sample value corresponding to each feature in each insurance data sample. The first determining unit is used to select any sample insurance data for any feature, determine the importance score of the feature in the corresponding sample insurance data based on the feature weight value corresponding to each feature and the sample value of each feature in the sample insurance data, and traverse all sample insurance data to obtain the importance score of the feature in each sample insurance data. The second determining unit is used to determine the importance score of the feature based on the importance score of the feature in each sample of insurance data; The traversal unit is used to traverse all features and determine the importance score for each feature.
[0073] Optionally, the second determining unit includes: The clustering subunit is used to cluster the importance scores of features in all sample insurance data to determine the initial sample insurance data. The first determining subunit is used to determine the difference between the sample value of the feature in each initial sample insurance data and the preset feature threshold. When the difference is greater than the preset difference threshold, the corresponding initial sample insurance data is determined as the target sample insurance data. The second determining subunit is used to determine the average importance score of the corresponding feature in all target sample insurance data as the importance score of the corresponding feature.
[0074] Optionally, the classification module 34 includes: The third determining unit is used to determine the first weight value of the first underwriting risk probability value and the second weight value of the second underwriting risk probability value; The unit is used to perform a weighted summation of the first underwriting risk probability value and the second underwriting risk probability value based on the first weight value and the second weight value, so as to obtain the target underwriting risk probability value. The classification unit is used to classify the underwriting risk of the policyholder according to the target underwriting risk probability value and determine the underwriting risk result of the policyholder.
[0075] Optionally, the classification units include: The third determining subunit is used to determine the clustering interval of the underwriting risk level, and the clustering interval of the underwriting risk level corresponds to the underwriting risk result; The fourth determination subunit is used to determine the target clustering interval corresponding to the target underwriting risk probability value based on the clustering interval of the underwriting risk level, and to determine the underwriting risk result corresponding to the target clustering interval based on the target clustering interval.
[0076] Optionally, the underwriting risk assessment device 30 also includes: The matching module is used to determine the wording that matches the underwriting risk results.
[0077] It should be noted that the information interaction and execution process between the above-mentioned units are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, which will not be repeated here.
[0078] Figure 4 This is a schematic diagram of the structure of a computer device provided in one embodiment of this application. For example... Figure 4 As shown, the computer device of this embodiment includes: at least one processor ( Figure 4 Only one is shown in the diagram), a memory, and a computer program stored in the memory and capable of running on at least one processor, which, when executed by the processor, implements the steps in any of the above-described underwriting risk assessment method embodiments.
[0079] This computer device may include, but is not limited to, a processor and memory. Those skilled in the art will understand that... Figure 4 The examples of computer devices are merely examples and do not constitute a limitation on computer devices. Computer devices may include more or fewer components than shown in the illustration, or combinations of certain components, or different components, such as network interfaces, displays, and input devices.
[0080] The processor referred to can be a CPU, but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor.
[0081] Memory includes readable storage media, internal memory, etc., wherein internal memory can be the RAM of a computer device, providing an environment for the operation of the operating system and computer-readable instructions stored in the readable storage media. The readable storage media can be the hard drive of a computer device, or in other embodiments, it can be an external storage device of the computer device, such as a plug-in hard drive, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card. Furthermore, memory can include both internal storage units and external storage devices of the computer device. Memory is used to store the operating system, applications, bootloader, data, and other programs, such as program code for computer programs. Memory can also be used to temporarily store data that has been output or will be output.
[0082] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above device can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here. If the integrated unit is implemented as 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, all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the above method embodiments. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. A computer-readable medium can include at least: any entity or device capable of carrying computer program code, a recording medium, a computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electrical carrier signals or telecommunication signals.
[0083] The implementation of all or part of the processes in the methods of the above embodiments can also be accomplished by a computer program product. When the computer program product is run on a computer device, it enables the computer device to execute the steps in the above method embodiments.
[0084] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.
[0085] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.
[0086] In the embodiments provided in this application, it should be understood that the disclosed apparatus / computer devices and methods can be implemented in other ways. For example, the apparatus / computer device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.
[0087] The units described as separate components may or may not be physically separate. 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 the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0088] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.
Claims
1. A method for underwriting risk assessment, characterized in that, The underwriting risk assessment methods include: Obtain the policyholder's insurance application data, process the insurance application data based on the first neural network model, and determine multiple features of the insurance application data and the feature weight value corresponding to each feature; Determine the importance score for each feature, and calculate the first underwriting risk probability value for the policyholder based on the feature weight value and the corresponding importance score for each feature. The insurance data is processed based on the second neural network model to determine the second underwriting risk probability value of the policyholder; Based on the first underwriting risk probability value and the second underwriting risk probability value, the underwriting risk of the policyholder is classified into different levels, and the underwriting risk result of the policyholder is determined.
2. The underwriting risk assessment method as described in claim 1, characterized in that, The determination of the importance score for each feature includes: Obtain training samples, which include insurance application data of multiple sample policyholders, and determine the sample value corresponding to each feature in each sample insurance application data; For any feature, select any sample insurance data, and determine the importance score of the feature in the corresponding sample insurance data based on the feature weight value corresponding to each feature and the sample value of each feature in the sample insurance data. Iterate through all sample insurance data to obtain the importance score of the feature in each sample insurance data. The importance score of the feature is determined based on its importance score in each sample of insurance data; Iterate through all features and determine the importance score for each feature.
3. The underwriting risk assessment method as described in claim 2, characterized in that, The step of determining the importance score of the feature based on its importance score in each sample of insurance data includes: Cluster the importance scores of the aforementioned features across all sample insurance data to determine the initial sample insurance data; The difference between the sample value of the feature in each initial sample insurance data and a preset feature threshold is determined. When the difference is greater than the preset difference threshold, the corresponding initial sample insurance data is determined as the target sample insurance data. The mean of the importance scores of the corresponding features in all the insurance data of the target samples is determined as the importance score of the corresponding feature.
4. The underwriting risk assessment method as described in claim 1, characterized in that, The step of classifying the underwriting risk of the policyholder according to the first underwriting risk probability and the second underwriting risk probability, and determining the underwriting risk result of the policyholder, includes: Determine a first weight value for the first underwriting risk probability value and a second weight value for the second underwriting risk probability value; Based on the first weight value and the second weight value, the first underwriting risk probability value and the second underwriting risk probability value are weighted and summed to obtain the target underwriting risk probability value; Based on the target underwriting risk probability value, the underwriting risk of the policyholder is classified into different levels to determine the underwriting risk result of the policyholder.
5. The underwriting risk assessment method as described in claim 4, characterized in that, The step of classifying the underwriting risk of the policyholder according to the target underwriting risk probability value and determining the underwriting risk result of the policyholder includes: Determine the clustering intervals for underwriting risk levels, wherein the clustering intervals for underwriting risk levels correspond to the underwriting risk results; Based on the clustering interval of the underwriting risk level, determine the target clustering interval corresponding to the target underwriting risk probability value, and based on the target clustering interval, determine the underwriting risk result corresponding to the target clustering interval.
6. The underwriting risk assessment method as described in claim 1, characterized in that, After determining the underwriting risk result of the policyholder, the following is also included: Based on the underwriting risk results, determine the wording that matches the underwriting risk results.
7. An underwriting risk assessment device, characterized in that, The underwriting risk assessment device includes: The first determining module is used to acquire the policyholder's insurance data, process the insurance data based on the first neural network model, and determine multiple features of the insurance data and the feature weight value corresponding to each feature; The calculation module is used to determine the importance score of each feature, and calculate the first underwriting risk probability value of the policyholder based on the feature weight value and the corresponding importance score of each feature. The second determining module is used to process the insurance application data based on the second neural network model to determine the second underwriting risk probability value of the policyholder; The classification module is used to classify the underwriting risk of the policyholder according to the first underwriting risk probability value and the second underwriting risk probability value, and to determine the underwriting risk result of the policyholder.
8. The underwriting risk assessment device as described in claim 7, characterized in that, The computing module includes: An acquisition unit is used to acquire training samples, which include insurance data of multiple sample policyholders and determine the sample value corresponding to each feature in each sample insurance data. The first determining unit is used to select any sample insurance data for any feature, determine the importance score of the feature in the corresponding sample insurance data according to the feature weight value corresponding to each feature and the sample value of each feature in the sample insurance data, and traverse all sample insurance data to obtain the importance score of the feature in each sample insurance data. The second determining unit is used to determine the importance score of the feature based on the importance score of the feature in each sample of insurance data; The traversal unit is used to traverse all features and determine the importance score for each feature.
9. A computer device, characterized in that, The computer device includes a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the underwriting risk assessment method as described in any one of claims 1 to 6.
10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the underwriting risk assessment method as described in any one of claims 1 to 6.