Risk prediction method and device, equipment and storage medium

Abstract

The invention provides a risk prediction method and device, equipment and a storage medium, and can be applied to the technical field of deep learning. The method comprises the steps that insuring data related to an insuring object is acquired, and the insuring data comprises target insurance policy data and health data of the insuring object, historical transaction insurance policy risk level data of a target agent object related to the insuring object, data of a mechanism where the target agent object is located and regional data; processing the insurance data according to a binning function to obtain key features; and inputting the key features into a risk prediction model, and outputting a risk prediction result.

Description

Technical Field

[0001] The present disclosure relates to the field of deep learning technology, and more specifically, to a risk prediction method, device, equipment and storage medium. Background Art

[0002] Underwriting risk control is an important part of the insurance industry. Its main task is to review and assess the risks of the insured's insurance application and determine the underwriting conditions and premium levels. Underwriting occupies an important position in the business process of insurance companies. Its accuracy and efficiency directly affect the risk control and operating results of insurance companies. The existing automatic underwriting rules generally involve filling out intelligent questionnaires online. The system automatically assesses risks based on the customer's answers. The risk assessment accuracy is low and it relies too much on manual underwriting, resulting in high labor costs. Summary of the invention

[0003] In view of this, the present disclosure provides a risk prediction method, apparatus, device and storage medium.

[0004] One aspect of the present disclosure provides a risk prediction method, comprising:

[0005] Obtain insurance data related to the insured object, wherein the insurance data includes the target policy data and health data of the insured object, as well as the historical transaction policy risk level data of the target agent object related to the insured object, the institution data of the target agent object, and the regional data; process the insurance data according to the binning function to obtain key features; input the key features into the risk prediction model, and output the risk prediction results.

[0006] According to an embodiment of the present disclosure, the insured object represents a target high-risk object selected based on a risk screening rule, and the insured object includes at least one of the following:

[0007] Historical critical illness insurance purchasing objects, historical target claims resource objects, and historical target insurance rate objects.

[0008] According to an embodiment of the present disclosure, the insurance data includes variable attributes and variable attribute values; wherein the insurance data is processed according to the binning function to obtain key features, including:

[0009] Based on the binning function, the insured data is grouped to obtain the risk assessment values ​​corresponding to the variable attributes in multiple grouping dimensions; for the variable attributes, according to the number of sub-positive samples and sub-negative samples corresponding to each grouping dimension, the positive sample ratio and negative sample ratio corresponding to each grouping dimension are obtained; for each grouping dimension of the variable attributes, according to the positive sample ratio, negative sample ratio and risk assessment value corresponding to each grouping dimension, the sub-information value corresponding to each grouping dimension is obtained; according to the sum of the sub-information values ​​corresponding to each grouping dimension, the information value of the variable attributes is obtained; when the information value of the variable attributes is greater than the preset threshold, the variable attributes are determined as key features.

[0010] According to an embodiment of the present disclosure, based on the binning function, the insurance data is grouped to obtain risk assessment values ​​corresponding to the variable attributes in multiple grouping dimensions, including:

[0011] Based on the preset grouping rules, each variable attribute in the insured data is grouped to obtain the grouping dimension corresponding to each variable attribute; according to the total number of positive samples corresponding to the variable attribute and the number of sub-positive samples corresponding to each grouping dimension, the proportion of positive samples corresponding to each grouping dimension is obtained; according to the total number of negative samples corresponding to the variable attribute and the number of sub-negative samples corresponding to each grouping dimension, the proportion of negative samples corresponding to each grouping dimension is obtained; according to the proportion of positive samples and the proportion of negative samples corresponding to each grouping dimension, the risk assessment value corresponding to each grouping dimension is determined.

[0012] According to an embodiment of the present disclosure, the risk prediction model is constructed based on a gradient boosting tree algorithm, a decision tree algorithm and a logistic regression algorithm; wherein key features are input into the risk prediction model, and a risk prediction result is output, including:

[0013] The key features are processed according to the gradient boosting tree algorithm to obtain gradient features; the gradient features are processed according to the decision tree to obtain target features; the target features are processed according to the logistic regression algorithm to obtain risk prediction results.

[0014] According to an embodiment of the present disclosure, the risk prediction model is trained by the following method:

[0015] Acquire training samples, the training samples include sample insurance policy data and classification labels related to the insured object sample, wherein the sample insurance data include historical insurance policy data and historical health data of the insured object sample, as well as historical transaction policy risk level data of historical agents related to the insured object sample, institutional data of historical agents, and regional data, and the classification labels represent the risk results of the insured object sample; process the sample insurance data according to the binning function to obtain the sample key features; input the sample key features into the initial risk prediction model, and output the sample risk prediction results; adjust the model parameters of the initial risk prediction model according to the sample risk prediction results and the classification labels to obtain the trained risk prediction model.

[0016] According to an embodiment of the present disclosure, an initial risk prediction model is trained based on the sample risk prediction results and classification labels to obtain a trained risk prediction model, including:

[0017] The sample risk prediction results and classification labels are processed according to the binary classification loss function to obtain the loss value; based on the loss value, the initial risk prediction model is trained to obtain the trained risk prediction model.

[0018] Another aspect of the present disclosure provides a risk prediction device, comprising:

[0019] An acquisition module is used to acquire insurance data related to the insured object, wherein the insurance data includes the target policy data and health data of the insured object, as well as the historical transaction policy risk level data of the target agent object related to the insured object, the target agent object's institution data, and the regional data;

[0020] The processing module is used to process the insurance data according to the binning function to obtain key features;

[0021] The generation module is used to input key features into the risk prediction model and output the risk prediction results.

[0022] Another aspect of the present disclosure provides an electronic device, comprising:

[0023] one or more processors;

[0024] a memory for storing one or more programs,

[0025] When one or more programs are executed by one or more processors, the one or more processors implement the risk prediction method as described above.

[0026] Another aspect of the present disclosure provides a computer-readable storage medium storing computer-executable instructions, which are used to implement the above risk prediction method when executed.

[0027] According to the risk prediction method, device, equipment and storage medium provided by the present disclosure, the insurance data is processed according to the binning function to obtain key features; the key features are input into the risk prediction model to output the risk prediction results. Since the insurance data includes the target policy data and health data of the insured object, as well as the historical transaction policy risk level data of the target agent object related to the insured object, the institution data of the target agent object, and the regional data, the insurance data can cover multi-dimensional information related to the insured object, and then the binning function can be used to extract deeper key features, thereby improving the prediction accuracy of the risk prediction model in assessing risks, reducing the reliance on manual underwriting, and thus reducing labor costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The above and other objects, features and advantages of the present disclosure will become more apparent through the following description of the embodiments of the present disclosure with reference to the accompanying drawings, in which:

[0029] Figure 1 The application scenario of the risk prediction method according to the embodiment of the present disclosure is schematically shown;

[0030] Figure 2 A flow chart of a risk prediction method according to an embodiment of the present disclosure is schematically shown;

[0031] Figure 3 A flowchart of training an initial risk prediction model according to an embodiment of the present disclosure is schematically shown;

[0032] Figure 4 A block diagram schematically shows a risk prediction device according to an embodiment of the present disclosure; and

[0033] Figure 5 A block diagram of an electronic device suitable for implementing a risk prediction method according to an embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0034] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present disclosure. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0035] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise", "include", etc. used herein indicate the existence of features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0036] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0037] When using expressions such as "at least one of A, B, and C, etc.", they should generally be interpreted according to the meaning of the expression commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0038] In the process of realizing the present disclosure, it was found that underwriting risk control is an important link in the insurance industry. The main task is to review and assess the risk of the insured's insurance application and determine the underwriting conditions and premium levels. Underwriting occupies an important position in the business process of insurance companies. Its accuracy and efficiency directly affect the risk control and operating results of insurance companies. The existing automatic underwriting rules generally fill out an intelligent questionnaire online. The system automatically assesses risks based on the customer's answers. The risk assessment accuracy is low, and it relies too much on manual underwriting, resulting in high labor costs.

[0039] In view of this, the embodiments of the present disclosure provide a risk prediction method, device, equipment and storage medium. The method includes: obtaining insurance data related to the insured object, wherein the insurance data includes the target policy data and health data of the insured object, and the historical transaction policy risk level data of the target agent object related to the insured object, the institution data of the target agent object, and the regional data; processing the insurance data according to the binning function to obtain key features; inputting the key features into the risk prediction model, and outputting the risk prediction results.

[0040] In the technical solution of the present disclosure, the user information (including but not limited to user personal information, user image information, user device information, such as location information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved are all information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of the relevant data comply with relevant laws, regulations and standards, take necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0041] Figure 1 The application scenario of the risk prediction method according to the embodiment of the present disclosure is schematically illustrated.

[0042] like Figure 1 As shown, the system architecture 100 according to this embodiment may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104 and a server 105. The network 104 is used to provide a medium for a communication link between the first terminal device 101, the second terminal device 102, the third terminal device 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0043] The user may use at least one of the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 through the network 104 to receive or send messages, etc. Various communication client applications may be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc. (only for example).

[0044] The first terminal device 101, the second terminal device 102, and the third terminal device 103 may be various electronic devices having display screens and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, and the like.

[0045] The server 105 may be a server that provides various services, such as a background management server (only as an example) that provides support for websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103. The background management server may analyze and process the received data such as user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal device.

[0046] It should be noted that the risk prediction method provided in the embodiment of the present disclosure can generally be executed by the server 105. Accordingly, the risk prediction device provided in the embodiment of the present disclosure can generally be set in the server 105. The risk prediction method provided in the embodiment of the present disclosure can also be executed by a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105. Accordingly, the risk prediction device provided in the embodiment of the present disclosure can also be set in a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105.

[0047] Alternatively, the risk prediction method provided in the embodiment of the present disclosure may also be executed by the first terminal device 101, the second terminal device 102 or the third terminal device 103, or may also be executed by other terminal devices different from the first terminal device 101, the second terminal device 102 or the third terminal device 103. Accordingly, the risk prediction device provided in the embodiment of the present disclosure may also be provided in the first terminal device 101, the second terminal device 102 or the third terminal device 103, or may be provided in other terminal devices different from the first terminal device 101, the second terminal device 102 or the third terminal device 103.

[0048] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0049] Figure 2 The flowchart of the risk prediction method according to the embodiment of the present disclosure is schematically shown.

[0050] like Figure 2 As shown, the method 200 includes operations S210 to S230.

[0051] In operation S210, insurance data related to the insured object is acquired.

[0052] According to an embodiment of the present disclosure, the insurance data includes the target policy data and health data of the insured object, as well as the historical transaction policy risk level data of the target agent object related to the insured object, the institution data and regional data of the target agent object.

[0053] According to an embodiment of the present disclosure, the insured is a customer who submits an insurance application to the insurance system through an online or offline channel. The target policy data of the insured represents the policy data in the current insurance application, and the target policy data may include information such as the insured's age, gender, family members, and insurance type.

[0054] According to an embodiment of the present disclosure, the health data may come from the physical examination report results of the insured person, and the health data may include indicator data such as the heart rate and blood pressure of the insured person.

[0055] According to an embodiment of the present disclosure, the target agent object related to the insured object represents the agent object that guides the insured object to submit the current insurance application. The historical transaction policy risk level data represents the risk level data of successful historical policy transactions, such as heavy risk, medium risk, etc.

[0056] According to an embodiment of the present disclosure, the historical transaction policy risk level data may also include data such as the claim rate, rejection rate, extension rate, and insurance types of the historical transaction policies of the target agent object.

[0057] According to the embodiments of the present disclosure, the institutional data and regional data of the target agent object can reflect the quality and coverage rate of the current insurance application.

[0058] In operation S220, the insurance data is processed according to the binning function to obtain key features.

[0059] According to an embodiment of the present disclosure, binning the insurance data according to the binning function may include: first cleaning, standardizing and normalizing the insurance data, converting the text feature data into numerical feature data, and then binning the continuous numerical feature data, so as to obtain key features with strong robustness.

[0060] For example, key features may include the insured person's age, gender, medical history, family medical history, target agent's agent quality, and other features.

[0061] In operation S230, the key features are input into the risk prediction model, and the risk prediction results are output.

[0062] According to the embodiments of the present disclosure, the key features obtained according to the binning function improve the robustness and prevent the overfitting of the risk prediction model.

[0063] According to an embodiment of the present disclosure, key features are input into a risk prediction model, and a risk prediction result is output, wherein the risk prediction result represents a risk score result of the insured's current insurance application.

[0064] According to an embodiment of the present disclosure, the risk prediction result may be a high-risk result or a risk-pass result. The risk prediction result provides decision support for underwriters. For example, if a risk-pass result is obtained, the risk prediction result is returned to the client where the insured is located and the insured is reminded to continue to complete the insurance process; if a high-risk result is obtained, a physical examination document or insurance rejection information is sent to the client where the insured is located.

[0065] According to the embodiments of the present disclosure, the insurance data includes the target policy data and health data of the insured object, as well as the historical transaction policy risk level data of the target agent object related to the insured object, the institutional data and regional data of the target agent object. Since the insurance data can cover multi-dimensional information related to the insured object, deeper key features can be extracted through the binning function, which improves the accuracy of the risk prediction model in assessing risks, reduces the reliance on manual underwriting, and thereby reduces labor costs.

[0066] According to an embodiment of the present disclosure, the insured object represents a target high-risk object selected based on a risk screening rule, and the insured object includes at least one of the following:

[0067] Historical critical illness insurance purchasing objects, historical target claims resource objects, and historical target insurance rate objects.

[0068] According to an embodiment of the present disclosure, the risk screening rules are aimed at selecting high-risk objects, the insurance system selects insured objects based on the risk screening rules and historical policy transaction data, and the target agent object uses the insured objects selected in the insurance system as target high-risk objects.

[0069] According to an embodiment of the present disclosure, the historical critical illness insurance purchase objects are objects that have purchased critical illness insurance policies in historical policy transactions, and the historical critical illness insurance purchase objects include long-term critical illness insurance purchase objects and short-term critical illness insurance purchase objects.

[0070] According to an embodiment of the present disclosure, the historical target claim resource object is an object that exceeds the claim resource threshold in historical policy transactions. For example, an object in medical insurance with a claim amount exceeding 30,000 yuan within two years is a historical target claim resource object.

[0071] According to an embodiment of the present disclosure, the historical target insurance rate object is an object that exceeds the insurance rate threshold in historical policy transactions. For example, an object with an insurance frequency exceeding 5 times is a historical target insurance rate object.

[0072] According to the embodiments of the present disclosure, the insured object focuses on a specific target high-risk object group, which improves the accurate target positioning and enables the risk prediction model to effectively identify and manage the high-risk object group.

[0073] According to an embodiment of the present disclosure, the insurance data includes variable attributes and variable attribute values; wherein the insurance data is processed according to the binning function to obtain key features, including:

[0074] Based on the binning function, the insured data is grouped to obtain the risk assessment values ​​corresponding to the variable attributes in multiple grouping dimensions; for the variable attributes, according to the number of sub-positive samples and sub-negative samples corresponding to each grouping dimension, the positive sample ratio and negative sample ratio corresponding to each grouping dimension are obtained; for each grouping dimension of the variable attributes, according to the positive sample ratio, negative sample ratio and risk assessment value corresponding to each grouping dimension, the sub-information value corresponding to each grouping dimension is obtained; according to the sum of the sub-information values ​​corresponding to each grouping dimension, the information value of the variable attributes is obtained; when the information value of the variable attributes is greater than the preset threshold, the variable attributes are determined as key features.

[0075] According to an embodiment of the present disclosure, variable attributes include multiple variable attributes, and the variable attributes may be attributes such as gender and age. The variable attributes correspond to variable attribute values. For example, the variable attribute value of gender may be "female".

[0076] According to the embodiment of the present disclosure, the text variable attributes in the insurance data are converted into numerical values, and then each variable attribute in the insurance data is grouped according to the binning function, and the risk assessment values ​​corresponding to the variable attributes in multiple grouping dimensions are obtained according to the variable attribute values ​​and the grouping rules. The risk assessment value can be calculated according to the weight of evidence (WOE).

[0077] According to an embodiment of the present disclosure, the insurance data includes data of multiple insured objects, and a risk attribute is selected from the insurance data. The risk attribute value corresponding to each insured object is 1 or 0. For the i-th grouping dimension of a variable attribute in the insurance data, when the risk attribute value corresponding to the insured object is 1, the variable attribute value corresponding to the insured object in the insurance data is a sub-positive sample; when the risk attribute value corresponding to the insured object is 0, the variable attribute value corresponding to the insured object in the insurance data is a sub-negative sample.

[0078] According to an embodiment of the present disclosure, for a certain variable attribute in the insurance data, this variable attribute corresponds to the variable attribute values ​​of different insured objects, and the sum of the number of sub-positive samples and the number of sub-negative samples on the i-th grouping dimension is the total number of variable attribute values ​​on the i-th grouping dimension.

[0079] According to an embodiment of the present disclosure, the proportion of positive samples on the i-th grouping dimension is obtained according to the ratio between the number of sub-positive samples on the i-th grouping dimension and the total number of variable attribute values ​​on the i-th grouping dimension; the proportion of negative samples on the i-th grouping dimension is obtained according to the ratio between the number of sub-negative samples on the i-th grouping dimension and the total number of variable attribute values ​​on the i-th grouping dimension.

[0080] According to an embodiment of the present disclosure, the positive sample ratio, negative sample ratio and risk assessment value corresponding to each grouping dimension are processed according to the information value function to obtain the sub-information value (Information Value, iv) corresponding to each grouping dimension, and then the sub-information value of the variable attribute in each grouping dimension is accumulated to obtain the information value of the variable attribute.

[0081] According to an embodiment of the present disclosure, the preset threshold may be 0.5, and the variable attribute whose information value is greater than the preset threshold is taken as a key feature. The key feature represents a feature with high risk value for assessment.

[0082] According to an embodiment of the present disclosure, based on the binning function, the insurance data is grouped to obtain risk assessment values ​​corresponding to the variable attributes in multiple grouping dimensions, including:

[0083] Based on the preset grouping rules, each variable attribute in the insured data is grouped to obtain the grouping dimension corresponding to each variable attribute; according to the total number of positive samples corresponding to the variable attribute and the number of sub-positive samples corresponding to each grouping dimension, the proportion of positive samples corresponding to each grouping dimension is obtained; according to the total number of negative samples corresponding to the variable attribute and the number of sub-negative samples corresponding to each grouping dimension, the proportion of negative samples corresponding to each grouping dimension is obtained; according to the proportion of positive samples and the proportion of negative samples corresponding to each grouping dimension, the risk assessment value corresponding to each grouping dimension is determined.

[0084] According to an embodiment of the present disclosure, the preset grouping rule may be division according to interval distance or division according to frequency.

[0085] For example, the division according to interval distance may include: taking the difference between the maximum variable attribute value and the minimum variable attribute value corresponding to the variable attribute as the interval distance, and then dividing the interval distance into four grouping dimensions, and the interval difference of each grouping dimension is the same.

[0086] According to an embodiment of the present disclosure, the total number of positive samples of the variable attribute is obtained according to the sum of the number of sub-positive samples of the variable attribute in multiple grouping dimensions in the insurance data. The proportion of positive samples corresponding to each grouping dimension is obtained according to the ratio of the number of sub-positive samples corresponding to each grouping dimension to the total number of positive samples.

[0087] According to an embodiment of the present disclosure, the total number of negative samples of the variable attribute is obtained according to the sum of the number of sub-negative samples of the variable attribute in multiple grouping dimensions in the insurance data. The proportion of negative samples corresponding to each grouping dimension is obtained according to the ratio of the number of sub-negative samples corresponding to each grouping dimension to the total number of negative samples.

[0088] According to an embodiment of the present disclosure, the positive sample proportion and negative sample proportion corresponding to each grouping dimension are processed according to a weight function to obtain a risk assessment value corresponding to each grouping dimension. The larger the risk assessment value, the larger the proportion of negative samples.

[0089] According to an embodiment of the present disclosure, the risk prediction model is constructed based on a gradient boosting tree algorithm, a decision tree algorithm and a logistic regression algorithm; wherein key features are input into the risk prediction model, and a risk prediction result is output, including:

[0090] The key features are processed according to the gradient boosting tree algorithm to obtain gradient features; the gradient features are processed according to the decision tree to obtain target features; the target features are processed according to the logistic regression algorithm to obtain risk prediction results.

[0091] According to an embodiment of the present disclosure, the risk prediction model is constructed based on an integrated machine learning algorithm (eXtremeGradient Boosting, XGB) of a decision tree, and can also be constructed based on a fusion of a gradient boosting tree algorithm, a decision tree algorithm, and a logistic regression algorithm.

[0092] According to an embodiment of the present disclosure, the Gradient Boosted Decision Trees (GBDT) algorithm is a boosting algorithm. According to the Gradient Boosted Trees algorithm, key features are processed, and the negative gradient of the cost function is used to fit the reduction of the cost function in each round of iteration to obtain gradient features. The Gradient Boosted Trees algorithm improves the stability and accuracy of the risk prediction model through symmetric trees and boundary processing.

[0093] According to the embodiments of the present disclosure, the gradient features are processed according to a decision tree to obtain the target features; the target features are processed according to a logistic regression algorithm to obtain the risk prediction results.

[0094] According to an embodiment of the present disclosure, a risk prediction result can be obtained by processing key features according to a risk prediction model constructed based on an integrated machine learning algorithm based on a decision tree.

[0095] According to the embodiments of the present disclosure, the decision tree-based integrated machine learning algorithm can accurately capture the complex relationships in key features by constructing an integration of decision trees, thereby providing highly accurate risk prediction results for underwriting risk control; through the design of parallel fusion of multiple algorithms such as gradient boosting tree algorithm, decision tree algorithm and logistic regression algorithm, not only the robustness of the risk prediction model is enhanced, but also the insured is provided with the ability to select the most appropriate algorithm according to specific insurance needs and insurance scenarios.

[0096] Figure 3 A flowchart for training an initial risk prediction model according to an embodiment of the present disclosure is schematically shown.

[0097] like Figure 3 As shown, the training method includes operations S310 to S340.

[0098] In step S310, a training sample is obtained.

[0099] In step S320, the sample insurance data is processed according to the binning function to obtain the sample key features.

[0100] In step S330, the key features of the sample are input into the initial risk prediction model, and the sample risk prediction results are output.

[0101] In step S340, the model parameters of the initial risk prediction model are adjusted according to the sample risk prediction results and classification labels to obtain a trained risk prediction model.

[0102] According to an embodiment of the present disclosure, the training samples include sample insurance policy data and classification labels related to the insured object sample, wherein the sample insurance data includes historical insurance policy data, historical health data of the insured object sample, and historical transaction policy risk level data of historical agent objects related to the insured object sample, institutional data of the historical agent objects, and regional data, and the classification labels represent the risk results of the insured object sample.

[0103] According to an embodiment of the present disclosure, sample policy data of a sample of insured persons is obtained from an insurance system. The historical policy data of the sample of insured persons represents the policy data in historical insurance applications, and the historical health data may come from historical physical examination report results of the sample of insured persons.

[0104] According to an embodiment of the present disclosure, the historical proxy object related to the insured object sample represents the proxy object that guides the insured object sample to submit a historical insurance application. For example, the historical proxy object's institution data is x1 branch institution, and the region data is xx city.

[0105] According to an embodiment of the present disclosure, the classification labels may be high-risk results and risk-pass results.

[0106] According to an embodiment of the present disclosure, sample insurance data is processed according to a binning function to obtain sample key features, which may include age, gender, medical history, family medical history, target agent object agent quality and other features of the insured object sample.

[0107] According to an embodiment of the present disclosure, the sample risk prediction results and classification labels can be processed according to an adaptive optimizer to obtain model performance, and then the model parameters of the initial risk prediction model can be adjusted according to the model performance, and the training is iterated until the model performance meets the preset performance value.

[0108] According to the embodiments of the present disclosure, binning optimization is performed based on multi-dimensional data such as historical policy data, historical health data, historical transaction policy risk level data, historical agency object institution data, and regional data to obtain deep sample key features, and then the initial risk prediction model is trained based on the sample key features to accurately capture the changes in the sample risks of the insured objects, thereby improving the accuracy of the risk prediction model.

[0109] According to an embodiment of the present disclosure, an initial risk prediction model is trained based on the sample risk prediction results and classification labels to obtain a trained risk prediction model, including:

[0110] The sample risk prediction results and classification labels are processed according to the binary classification loss function to obtain the loss value; based on the loss value, the initial risk prediction model is trained to obtain the trained risk prediction model.

[0111] According to an embodiment of the present disclosure, the binary classification loss function may be a binary cross entropy loss function (BCELoss). The sample risk prediction results and classification labels of the insured object samples are processed according to the binary classification loss function to obtain a loss value.

[0112] According to an embodiment of the present disclosure, the model parameters of the initial risk prediction model are adjusted according to the loss value, and iterative training is performed until the model performance meets the preset loss value. For example, the preset loss value is 85%, and when the loss value is greater than 85%, iterative training is stopped.

[0113] Based on the above risk prediction method, the present disclosure also provides a risk prediction device. Figure 4 The device is described in detail.

[0114] Figure 4 A structural block diagram of a risk prediction device according to an embodiment of the present disclosure is shown.

[0115] like Figure 4 As shown, the risk prediction device 400 of this embodiment includes an acquisition module 410 , a processing module 420 and a generation module 430 .

[0116] The acquisition module 410 is used to acquire insurance data related to the insured object, wherein the insurance data includes the target policy data and health data of the insured object, as well as the historical transaction policy risk level data of the target agent object related to the insured object, the institution data of the target agent object, and the regional data. In one embodiment, the acquisition module 410 can be used to perform the operation S210 described above, which will not be repeated here.

[0117] The processing module 420 is used for processing the insurance data according to the binning function to obtain key features. In one embodiment, the processing 420 can be used to perform the operation S220 described above, which will not be repeated here.

[0118] The generation module 430 is used to input the key features into the risk prediction model and output the risk prediction result. In one embodiment, the generation module 430 can be used to perform the operation S230 described above, which will not be described in detail here.

[0119] According to an embodiment of the present disclosure, the processing module 420 includes a first processing submodule, a second processing submodule, a third processing submodule, a fourth processing submodule and a fifth processing submodule.

[0120] The first processing submodule is used to group the insurance data based on the binning function to obtain the risk assessment values ​​corresponding to the variable attributes in multiple grouping dimensions.

[0121] The second processing submodule is used to obtain the positive sample ratio and the negative sample ratio corresponding to each grouping dimension according to the number of sub-positive samples and the number of sub-negative samples corresponding to each grouping dimension based on the variable attributes.

[0122] The third processing submodule is used to obtain the sub-information value corresponding to each grouping dimension of the variable attribute according to the positive sample ratio, negative sample ratio and risk assessment value corresponding to each grouping dimension.

[0123] The fourth processing submodule is used to obtain the information value of the variable attribute according to the sum of the sub-information values ​​corresponding to each grouping dimension.

[0124] The fifth processing submodule is used to determine the variable attribute as a key feature when the information value of the variable attribute is greater than a preset threshold.

[0125] According to an embodiment of the present disclosure, the first processing submodule includes a first processing subunit, a second processing unit, a third processing unit, and a fourth processing unit.

[0126] The first processing sub-unit is used to group each variable attribute in the insurance data based on a preset grouping rule to obtain a grouping dimension corresponding to each variable attribute.

[0127] The second processing sub-unit is used to obtain the positive sample proportion corresponding to each grouping dimension according to the total positive sample quantity corresponding to the variable attribute and the sub-positive sample quantity corresponding to each grouping dimension.

[0128] The third processing subunit is used to obtain the negative sample proportion corresponding to each grouping dimension according to the total number of negative samples corresponding to the variable attribute and the number of sub-negative samples corresponding to each grouping dimension.

[0129] The fourth processing subunit is used to determine the risk assessment value corresponding to each grouping dimension according to the positive sample proportion and the negative sample proportion corresponding to each grouping dimension.

[0130] According to an embodiment of the present disclosure, the generation module 430 includes a first generation submodule, a second generation submodule and a third generation submodule.

[0131] The first generation submodule is used to process key features according to the gradient boosting tree algorithm to obtain gradient features.

[0132] The second generation submodule is used to process the gradient features according to the decision tree to obtain the target features.

[0133] The third generation submodule is used to process the target features according to the logistic regression algorithm to obtain the risk prediction results.

[0134] According to the embodiments of the present invention, any one or more of the modules, submodules, units, and subunits, or at least part of the functions of any one of them can be implemented in one module. According to the embodiments of the present invention, any one or more of the modules, submodules, units, and subunits can be split into multiple modules for implementation. According to the embodiments of the present invention, any one or more of the modules, submodules, units, and subunits can be at least partially implemented as hardware circuits, such as field programmable gate arrays (FPGAs), programmable logic arrays (PLAs), systems on chips, systems on substrates, systems on packages, application specific integrated circuits (ASICs), or can be implemented by hardware or firmware in any other reasonable way of integrating or packaging circuits, or can be implemented in any one of the three implementation methods of software, hardware, and firmware, or in any appropriate combination of any of them. Alternatively, according to the embodiments of the present invention, one or more of the modules, submodules, units, and subunits can be at least partially implemented as computer program modules, and when the computer program modules are run, the corresponding functions can be performed.

[0135] For example, any multiple of the acquisition module 410, the processing module 420, and the generation module 430 can be combined in one module / unit / sub-unit for implementation, or any one of the modules / units / sub-units can be split into multiple modules / units / sub-units. Alternatively, at least part of the functions of one or more of these modules / units / sub-units can be combined with at least part of the functions of other modules / units / sub-units and implemented in one module / unit / sub-unit. According to an embodiment of the present disclosure, at least one of the acquisition module 410, the processing module 420, and the generation module 430 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or implemented in any one of the three implementation methods of software, hardware, and firmware, or in any appropriate combination of any of them. Alternatively, at least one of the acquisition module 410 , the processing module 420 , and the generation module 430 may be at least partially implemented as a computer program module, and when the computer program module is executed, the corresponding function may be performed.

[0136] Figure 5 A block diagram of an electronic device suitable for implementing a risk prediction method according to an embodiment of the present invention is shown.

[0137] Figure 5The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0138] like Figure 5 As shown, the computer electronic device 500 according to an embodiment of the present invention includes a processor 501, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 502 or the program loaded from the storage part 509 to the random access memory (RAM) 503. The processor 501 may include, for example, a general-purpose microprocessor (such as a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (for example, an application-specific integrated circuit (ASIC)), etc. The processor 501 may also include an on-board memory for caching purposes. The processor 501 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present invention.

[0139] In RAM 503, various programs and data required for the operation of electronic device 500 are stored. Processor 501, ROM 502 and RAM 503 are connected to each other via bus 504. Processor 501 performs various operations of the method flow according to the embodiment of the present invention by executing the program in ROM 502 and / or RAM 503. It should be noted that the program can also be stored in one or more memories other than ROM 502 and RAM 503. Processor 501 can also perform various operations of the method flow according to the embodiment of the present invention by executing the program stored in one or more memories.

[0140] Optionally, the electronic device 500 may further include an input / output (I / O) interface 505, which is also connected to the bus 504. The electronic device 500 may further include one or more of the following components connected to the input / output (I / O) interface 505: an input portion 506 including a keyboard, a mouse, etc.; an output portion 507 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage portion 508 including a hard disk, etc.; and a communication portion 509 including a network interface card such as a LAN card, a modem, etc. The communication portion 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to the input / output (I / O) interface 505 as needed. A removable medium 511, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 510 as needed, so that a computer program read therefrom is installed into the storage portion 508 as needed.

[0141] Optionally, the method flow according to the embodiment of the present invention can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable storage medium, and the computer program contains a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 509, and / or installed from the removable medium 511. When the computer program is executed by the processor 501, the above-mentioned functions defined in the system of the embodiment of the present invention are executed. Optionally, the system, equipment, device, module, unit, etc. described above can be implemented by a computer program module.

[0142] The present invention also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiment; or may exist independently without being assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the risk prediction method according to the embodiment of the present invention is implemented.

[0143] Optionally, the computer-readable storage medium may be a non-volatile computer-readable storage medium. For example, it may include, but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, apparatus, or device.

[0144] For example, optionally, the computer-readable storage medium may include the ROM 502 and / or the RAM 503 described above and / or one or more memories other than the ROM 502 and the RAM 503 .

[0145] An embodiment of the present invention also includes a computer program product, which includes a computer program, which contains program code for executing the method provided by the embodiment of the present invention. When the computer program product runs on an electronic device, the program code is used to enable the electronic device to implement the risk prediction method provided by the embodiment of the present invention.

[0146] When the computer program is executed by the processor 501, the above functions defined in the system / device of the embodiment of the present invention are performed. Optionally, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0147] In one embodiment, the computer program may rely on tangible storage media such as optical storage devices, magnetic storage devices, etc. In another embodiment, the computer program may also be transmitted and distributed in the form of signals on a network medium, and downloaded and installed through the communication part 509, and / or installed from the removable medium 511. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0148] Optionally, the program code for executing the computer program provided by the embodiment of the present invention can be written in any combination of one or more programming languages, specifically, these computing programs can be implemented using high-level process and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, such as Java, C++, python, "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on the remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (e.g., using an Internet service provider to connect through the Internet).

[0149] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram may represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions. It can be understood by those skilled in the art that the features recorded in the various embodiments of the present disclosure can be combined and / or combined in a variety of ways, even if such a combination or combination is not explicitly recorded in the present disclosure. In particular, without departing from the spirit and teaching of the present disclosure, the features described in the various embodiments of the present disclosure may be combined and / or combined in a variety of ways. All of these combinations and / or combinations fall within the scope of the present disclosure.

[0150] The embodiments of the present disclosure are described above. However, these embodiments are only for illustrative purposes and are not intended to limit the scope of the present disclosure. Although the embodiments are described above, this does not mean that the measures in the various embodiments cannot be used in combination to advantage. Without departing from the scope of the present disclosure, those skilled in the art may make a variety of substitutions and modifications, which should all fall within the scope of the present disclosure.

Claims

1. A risk prediction method, characterized in that: The method comprises: Acquire insurance data related to the insured object, wherein the insurance data includes target policy data and health data of the insured object, historical transaction policy risk level data of the target agent object related to the insured object, target agent object institution data, and regional data; Processing the insurance data according to a binning function to obtain key features; The key features are input into the risk prediction model and the risk prediction results are output.

2. The method according to claim 1, characterized in that The insured object represents a target high-risk object selected based on risk screening rules, and the insured object includes at least one of the following: Historical critical illness insurance purchasing objects, historical target claims resource objects, and historical target insurance rate objects.

3. The method according to claim 1, characterized in that: The insurance data includes variable attributes and variable attribute values; The insurance data is processed according to the binning function to obtain key features, including: Based on the binning function, the insurance data is grouped to obtain the risk assessment values ​​corresponding to the variable attributes in multiple grouping dimensions; For the variable attributes, according to the number of sub-positive samples and the number of sub-negative samples corresponding to each grouping dimension, the positive sample ratio and the negative sample ratio corresponding to each grouping dimension are obtained; For each grouping dimension of the variable attribute, according to the positive sample ratio, the negative sample ratio and the risk assessment value corresponding to each grouping dimension, obtain the sub-information value corresponding to each grouping dimension; Obtaining the information value of the variable attribute according to the sum of the sub-information values ​​corresponding to each of the grouping dimensions; When the information value of the variable attribute is greater than a preset threshold, the variable attribute is determined as a key feature.

4. The method according to claim 3, characterized in that Based on the binning function, the insurance data is grouped to obtain the risk assessment values ​​corresponding to the variable attributes in multiple grouping dimensions, including: Based on a preset grouping rule, each variable attribute in the insurance data is grouped to obtain a grouping dimension corresponding to each variable attribute; According to the total number of positive samples corresponding to the variable attribute and the number of sub-positive samples corresponding to each grouping dimension, the proportion of positive samples corresponding to each grouping dimension is obtained; According to the total number of negative samples corresponding to the variable attribute and the number of sub-negative samples corresponding to each grouping dimension, the proportion of negative samples corresponding to each grouping dimension is obtained; According to the positive sample proportion and negative sample proportion corresponding to each grouping dimension, the risk assessment value corresponding to each grouping dimension is determined.

5. The method according to claim 1, characterized in that The risk prediction model is constructed based on the gradient boosting tree algorithm, the decision tree algorithm and the logistic regression algorithm; The key features are input into the risk prediction model, and the risk prediction results are output, including: Processing the key features according to the gradient boosting tree algorithm to obtain gradient features; Processing the gradient feature according to the decision tree to obtain a target feature; The target features are processed according to the logistic regression algorithm to obtain a risk prediction result.

6. The method according to claim 1, characterized in that The risk prediction model is trained by the following method: Acquire training samples, wherein the training samples include sample insurance policy data and classification labels related to the insured object sample, wherein the sample insurance data includes historical insurance policy data and historical health data of the insured object sample, and historical transaction insurance policy risk level data, historical agency object institution data, and regional data of historical agents related to the insured object sample, and the classification labels represent the risk results of the insured object sample; Processing the sample insurance data according to the binning function to obtain the sample key features; Inputting the sample key features into the initial risk prediction model and outputting the sample risk prediction results; According to the sample risk prediction results and the classification labels, the model parameters of the initial risk prediction model are adjusted to obtain a trained risk prediction model.

7. The method according to claim 6, characterized in that According to the sample risk prediction result and the classification label, the risk prediction model is trained to obtain a trained risk prediction model, including: Processing the sample risk prediction result and the classification label according to a binary classification loss function to obtain a loss value; The risk prediction model is trained according to the loss value to obtain a trained risk prediction model.

8. A risk prediction device, comprising: An acquisition module is used to acquire insurance data related to the insured object, wherein the insurance data includes the target policy data and health data of the insured object, as well as the historical transaction policy risk level data of the target agent object related to the insured object, the institution data of the target agent object, and the regional data; A processing module, used for processing the insurance data according to a binning function to obtain key features; The generation module is used to input the key features into the risk prediction model and output the risk prediction results.

9. An electronic device, comprising: one or more processors; a memory for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, enables the processor to implement the method according to any one of claims 1 to 7.

Images