Data generation method and device based on artificial intelligence, computer equipment and medium

Abstract

The invention belongs to the technical field of artificial intelligence, and relates to an artificial intelligence-based data generation method and device, computer equipment and a storage medium, and the method comprises the steps: obtaining enterprise data of a target enterprise; performing risk assessment on the enterprise data based on the risk assessment model to generate a risk score; calling a target generative adversarial network corresponding to a target industry of the target enterprise; generating an industry risk coefficient of the target industry based on the target generative adversarial network; generating an industry adjustment weight of the target industry based on the industry risk coefficient; generating an enterprise personality weight of the target enterprise based on a weight generation strategy; and performing calculation processing on the reference insurance premium, the risk score, the industry adjustment weight and the enterprise personality weight of the target enterprise based on an insurance premium calculation formula to obtain a target insurance premium of the target enterprise. In addition, the target insurance premium may be stored in the block chain. The premium pricing method and device can be applied to premium generation scenes in the field of financial insurance, and accurate premium pricing processing can be effectively achieved through the premium pricing method and device.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence, and can be applied to the field of financial technology, in particular to a data generation method and device based on artificial intelligence, a computer device and a storage medium. BACKGROUND

[0002] In the enterprise group insurance market, with the continuous improvement of the diversification and customization of enterprise customer demand, the traditional insurance pricing mode faces significant limitations. The current industry generally adopts a unified rate pricing mechanism, which only calculates the premium based on the overall attributes of the enterprise (such as the number of employees, industry category), and ignores the differentiated influence of individual risk characteristics of employees (such as age, gender, position level, health status and medical history, etc.). Such a rough pricing method leads to a mismatch between risk and premium, so that accurate premium pricing cannot be achieved, which easily leads to the situation that the high-risk employee group pays insufficient premium because the actual risk is not quantified, and the low-risk employee group bears excessive premium because of the dilution of the risk pool, thereby forming a typical adverse selection effect.

[0003] Therefore, it is urgent to provide a method that can achieve accurate premium pricing to achieve fair pricing while suppressing adverse selection, thereby improving the resource allocation efficiency and sustainable development capability of the group insurance market. SUMMARY

[0004] The purpose of the embodiments of the present application is to provide a data generation method and device based on artificial intelligence, a computer device and a storage medium, to solve the technical problem that the current unified rate pricing mechanism cannot achieve accurate premium pricing.

[0005] In a first aspect, a data generation method based on artificial intelligence is provided, comprising:

[0006] obtaining enterprise data of a target enterprise;

[0007] performing risk assessment processing on the enterprise data based on a preset risk assessment model to generate a corresponding risk score;

[0008] determining a target industry corresponding to the target enterprise, and calling a target generative adversarial network corresponding to the target industry;

[0009] generating an industry risk coefficient corresponding to the target industry based on the target generative adversarial network;

[0010] generating an industry adjustment weight corresponding to the target industry based on the industry risk coefficient;

[0011] generating an enterprise individual weight corresponding to the target enterprise based on a preset weight generation strategy;

[0012] The base premium of the target enterprise, the risk score, the industry adjustment weight, and the enterprise individual weight are calculated based on a preset premium calculation formula, to obtain a target premium of the target enterprise.

[0013] In a second aspect, a data generation device based on artificial intelligence is provided, comprising:

[0014] A first acquisition module is configured to acquire enterprise data of a target enterprise.

[0015] A first generation module is configured to perform risk assessment processing on the enterprise data based on a preset risk assessment model, to generate a corresponding risk score.

[0016] A first processing module is configured to determine a target industry corresponding to the target enterprise, and to call a target generative adversarial network corresponding to the target industry.

[0017] A second generation module is configured to generate an industry risk coefficient corresponding to the target industry based on the target generative adversarial network.

[0018] A third generation module is configured to generate an industry adjustment weight corresponding to the target industry based on the industry risk coefficient.

[0019] A fourth generation module is configured to generate an enterprise individual weight corresponding to the target enterprise based on a preset weight generation strategy.

[0020] A second processing module is configured to perform calculation processing on a base premium of the target enterprise, the risk score, the industry adjustment weight, and the enterprise individual weight based on a preset premium calculation formula, to obtain a target premium of the target enterprise.

[0021] In a third aspect, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned data generation method based on artificial intelligence when executing the computer program.

[0022] In a fourth aspect, a computer readable storage medium is provided, which stores a computer program, wherein the computer program is executed by a processor to implement the steps of the above-mentioned data generation method based on artificial intelligence.

[0023] The scheme implemented by the above artificial intelligence-based data generation method, device, computer equipment and storage medium first acquires enterprise data of a target enterprise; then performs risk assessment processing on the enterprise data based on a preset risk assessment model to generate a corresponding risk score; then determines a target industry corresponding to the target enterprise and calls a target generative adversarial network corresponding to the target industry; subsequently generates an industry risk coefficient corresponding to the target industry based on the target generative adversarial network; further generates an industry adjustment weight corresponding to the target industry based on the industry risk coefficient; generates an enterprise individual weight corresponding to the target enterprise based on a preset weight generation strategy; and finally calculates and processes the base premium of the target enterprise, the risk score, the industry adjustment weight and the enterprise individual weight based on a preset premium calculation formula to obtain a target premium of the target enterprise. After acquiring the enterprise data of the target enterprise, the present application generates a risk score by performing risk assessment processing on the enterprise data based on the use of a risk assessment model, generates an industry risk coefficient corresponding to the target industry based on the use of a target generative adversarial network corresponding to the target industry, generates an industry adjustment weight of the target industry based on the use of the industry risk coefficient, further generates an enterprise individual weight of the target enterprise based on the use of a weight generation strategy, and finally calculates and processes the base premium of the target enterprise, the risk score, the industry adjustment weight and the enterprise individual weight based on the use of a premium calculation formula, so that the target premium of the target enterprise can be automatically and accurately generated. The present application combines the multi-level factors corresponding to the base premium of the target enterprise, the risk score, the industry adjustment weight and the enterprise individual weight to perform premium calculation and processing, so that accurate, fair and reasonable premium pricing can be achieved, and the accuracy of the generated target premium is effectively improved. BRIEF DESCRIPTION OF DRAWINGS

[0024] In order to more clearly illustrate the schemes in the present application, the drawings needed in the description of the embodiments of the present application will be briefly introduced. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0025] Figure 1 is an exemplary system architecture diagram to which the present application can be applied;

[0026] Figure 2 is a flowchart of one embodiment of the artificial intelligence-based data generation method according to the present application;

[0027] Figure 3 is a structural schematic diagram of one embodiment of the artificial intelligence-based data generation device according to the present application;

[0028] Figure 4 is a structural schematic diagram of one embodiment of the computer device according to the present application. DETAILED DESCRIPTION

[0029] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs; the terminology used in the description herein is for describing particular embodiments only and is not intended to be limiting of the application; the use herein of terms such as "comprise" and "have" and any variations such as "comprises" and "has" is intended to cover the presence of successively stated integers or features but not preclude the presence of other integers or features; the use herein of terms such as "first", "second" and "third" and the like is intended to distinguish between different objects and not to describe a particular sequential order.

[0030] Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearances of the phrase "in an embodiment" in various places in the specification are not necessarily all referring to the same embodiment, nor are they necessarily mutually exclusive of one another.

[0031] In order to make the technical personnel in the art better understand the scheme of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings.

[0032] As shown in Figure 1 The system architecture 100 can include a terminal device 101, a network 102 and a server 103, and the terminal device 101 can be a notebook computer 1011, a tablet computer 1012 or a mobile phone 1013. The network 102 is a medium for providing a communication link between the terminal device 101 and the server 103. The network 102 can include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0033] A user can use the terminal device 101 to interact with the server 103 through the network 102 to receive or send messages, etc. Various communication client applications can be installed on the terminal device 101, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0034] The terminal device 101 can be various electronic devices with a display screen and supporting web browsing, in addition to the notebook computer 1011, the tablet computer 1012 or the mobile phone 1013, the terminal device 101 can also be an electronic book reader, an MP3 player (Moving Picture Experts Group Audio Layer III), an MP4 player (Moving Picture Experts Group Audio Layer IV), a laptop computer and a desktop computer, etc.

[0035] The server 103 can be a server providing various services, for example, a background server providing support for a page displayed on the terminal device 101.

[0036] It should be noted that the data generation method based on artificial intelligence provided by the embodiments of the present application is generally executed by a server / terminal device, and accordingly, the data generation apparatus based on artificial intelligence is generally arranged in a server / terminal device.

[0037] It should be understood that Figure 1 The number of terminal devices, networks and servers in

[0038] With reference to Figure 2 , a flow chart of one embodiment of the data generation method based on artificial intelligence according to the present application is shown. The order of the steps in the flow chart can be changed according to different needs, and some steps can be omitted. The data generation method based on artificial intelligence provided by the embodiments of the present application can be applied to any scenario requiring data generation, and then the data generation method based on artificial intelligence can be applied to products in these scenarios, for example, a premium generation scenario in the field of finance and insurance. The data generation method based on artificial intelligence comprises the following steps:

[0039] Step S201, obtaining enterprise data of a target enterprise.

[0040] In the present embodiment, the electronic device (for example Figure 1The server / terminal device shown) can obtain enterprise data of the target enterprise through wired connection or wireless connection. It should be noted that the wireless connection can include, but is not limited to, 3G / 4G / 5G connection, WiFi connection, Bluetooth connection, WiMAX connection, Zigbee connection, UWB (ultra wi deband) connection, and other now known or future developed wireless connection. The execution subject of the present application is a data generation system, also known as a premium generation system, which can be referred to as a system. The present application can be applied to the premium generation scene in the financial insurance field, that is, to price the premium for the enterprise. Exemplarily, the target enterprise can be a financial enterprise, such as an insurance enterprise, a bank, etc. The enterprise data includes the basic data and risk data of the target enterprise. The basic data can include industry code, employee list, enterprise size, region, etc. The risk data can include the work injury accident rate of the enterprise in recent years, the network security investment proportion, the abnormal rate of physical examination (employee health data), production safety record, etc. risk related data.

[0041] Specifically, for the industry code: obtain the standard industry classification code of the enterprise (such as manufacturing, finance, etc.) through enterprise registration information or industry association. For the employee list (including age information): collect the age distribution data of the enterprise employees, which can be obtained through the enterprise human resources department or employee management system. For enterprise size: determine the enterprise size (such as small, medium, large) according to the number of employees, annual turnover, etc. For regional information: record the region where the enterprise is located, which is used to consider the regional risk difference (such as areas with frequent natural disasters). For the work injury accident rate in recent years: extract the work injury accident rate in the past 3 years from the enterprise safety production record or work injury insurance claim data. For network security investment proportion: collect the proportion of annual investment in network security to total income of the enterprise, reflecting the importance of the enterprise to network risk. For the abnormal rate of physical examination: through the annual physical examination report of the enterprise employees, the proportion of abnormal indicators is calculated to evaluate the overall health status of the employees. For production safety record: collect safety violation records, accident reports, etc. in the production process of the enterprise to evaluate the production safety risk.

[0042] Step S202, performing risk assessment processing on the enterprise data based on a preset risk assessment model to generate a corresponding risk score.

[0043] In the present embodiment, the specific implementation process of the above-mentioned risk assessment processing on the enterprise data based on a preset risk assessment model to generate a corresponding risk score will be further described in detail in the subsequent specific embodiments, and will not be described here.

[0044] Step S203, determining a target industry corresponding to the target enterprise, and calling a target generative adversarial network corresponding to the target industry.

[0045] In this embodiment, the target industry to which the target enterprise belongs can be determined by obtaining the industry information of the target enterprise. A dedicated generative adversarial network is designed for different industries in advance to learn the distribution of industry-specific risk events. Specifically, the construction process of the industry-specific generative adversarial network includes: industry classification: different industries are classified according to industry codes (such as manufacturing industry, financial industry). Generator structure: an independent generative adversarial network (or generator) is designed for each industry, and the input is the industry code and historical risk event features (such as fault type, loss amount). The generator uses a deep neural network (such as a fully connected layer, a convolutional layer), and the output is simulated risk event data. Risk pattern learning: manufacturing industry generator: learn risk patterns such as equipment failure type (such as mechanical failure, electrical failure), downtime, repair cost, etc. Financial industry generator: learn risk patterns such as internal fraud type (such as corruption, data tampering), data leakage size, compliance violation record, etc. The trained generative adversarial network can generate a large amount of simulated risk scenario data.

[0046] In addition, the process of using the generative adversarial network for data processing includes: 1. Generating data: using the trained industry generator to generate a large amount of simulated risk scenario data. For example, the manufacturing industry generator generates data such as downtime caused by equipment failure, repair cost, etc. 2. Data discrimination: unified discriminator: design a cross-industry discriminator to determine whether the generated data conforms to the real risk distribution. Adversarial training: through the adversarial training of the generator and the discriminator, the quality of the generated data is improved. 3. Industry risk coefficient calculation: based on the risk scenario data generated by the generator, the industry risk coefficient is calculated. For example, the manufacturing industry risk coefficient can be obtained by weighting the equipment failure frequency, average loss amount, etc.

[0047] Step S204, generating an industry risk coefficient corresponding to the target industry based on the target generative adversarial network.

[0048] In this embodiment, the specific implementation process of generating an industry risk coefficient corresponding to the target industry based on the target generative adversarial network will be further described in detail in subsequent specific embodiments, and will not be described here.

[0049] Step S205, generating an industry adjustment weight corresponding to the target industry based on the industry risk coefficient.

[0050] In this embodiment, the specific implementation process of generating an industry adjustment weight corresponding to the target industry based on the industry risk coefficient will be further described in detail in subsequent specific embodiments, and will not be described here.

[0051] Step S206, generating the enterprise individual weight corresponding to the target enterprise based on the preset weight generation strategy.

[0052] In the embodiment, the specific implementation process of generating the enterprise individual weight corresponding to the target enterprise based on the preset weight generation strategy will be further described in details in subsequent specific embodiments, and will not be elaborated here.

[0053] Step S207, calculating the target premium of the target enterprise based on the preset premium calculation formula, the base premium, the risk score, the industry adjustment weight and the enterprise individual weight.

[0054] In the embodiment, the premium calculation formula is a multi-level dynamic premium formula, which specifically includes: Premium = Base × (1 + γi) × (1 + α × R + β × C). Wherein, Base is the base premium. γi is the industry adjustment weight. R is the risk score. C is the industry risk coefficient. (α, β) is the enterprise individual weight. The base premium, risk score, industry adjustment weight and enterprise individual weight of the target enterprise are respectively substituted into the premium calculation formula for calculation, and the calculation result is taken as the target premium of the target enterprise.

[0055] The application first acquires enterprise data of a target enterprise; then performs risk assessment processing on the enterprise data based on a preset risk assessment model to generate a corresponding risk score; then determines a target industry corresponding to the target enterprise and calls a target generative adversarial network corresponding to the target industry; subsequently generates an industry risk coefficient corresponding to the target industry based on the target generative adversarial network; further generates an industry adjustment weight corresponding to the target industry based on the industry risk coefficient; generates an enterprise individual weight corresponding to the target enterprise based on a preset weight generation strategy; and finally calculates and processes a benchmark premium of the target enterprise, the risk score, the industry adjustment weight, and the enterprise individual weight based on a preset premium calculation formula to obtain a target premium of the target enterprise. After acquiring the enterprise data of the target enterprise, the application generates a risk score by performing risk assessment processing on the enterprise data based on the use of the risk assessment model, generates an industry risk coefficient corresponding to the target industry based on the use of the target generative adversarial network corresponding to the target industry, generates an industry adjustment weight of the target industry based on the use of the industry risk coefficient, further generates an enterprise individual weight of the target enterprise based on the use of the weight generation strategy, and finally calculates and processes the benchmark premium, the risk score, the industry adjustment weight, and the enterprise individual weight of the target enterprise based on the use of the premium calculation formula, so as to automatically and accurately generate the target premium of the target enterprise. By combining the multi-level factors corresponding to the benchmark premium, the risk score, the industry adjustment weight, and the enterprise individual weight of the target enterprise, the application can accurately, fairly, and reasonably calculate and process the premium, effectively improving the accuracy of the generated target premium.

[0056] In some optional implementations, the enterprise data includes basic data and risk data; step S202 includes the following steps:

[0057] The basic data and the risk data are integrated to obtain corresponding feature data.

[0058] In this embodiment, the collected basic data and risk data are integrated into the form F of the input features of the risk assessment model, for example, F = [industry code (categorical variable, which needs to be encoded into a numerical form), employee age distribution (such as average age, age segment proportion), work injury accident rate in recent years (continuous variable), network security investment proportion (continuous variable), and other related features (such as physical examination abnormality rate, production safety record score, etc.)].

[0059] The feature data is preprocessed to obtain corresponding processing data.

[0060] In the embodiment, the feature data can be preprocessed in the same way as the training data of the risk assessment model to obtain the processing data, for example, the industry code is encoded into a numerical form, and the continuous variables such as age distribution and work accident rate are calculated to ensure that the format of the feature data is consistent with the training data.

[0061] The pre-constructed risk assessment model is called.

[0062] In the embodiment, the construction process of the risk assessment model will be further described in detail in subsequent specific embodiments, and will not be described here.

[0063] The risk assessment model is called.

[0064] In the embodiment, the processing data can be input into the trained risk assessment model, which will perform risk assessment on the input processing data and output a matching group risk score R. Score interpretation: R is a continuous value, reflecting the claim risk of the enterprise relative to the industry average. R reflects the claim risk of the enterprise relative to the industry average: score = 0.5: the risk of the enterprise is consistent with the industry average. Score> 0.5: the risk of the enterprise is higher than the industry average (such as 0.75 indicating high risk). Score < 0.5: the risk of the enterprise is lower than the industry average (such as 0.3 indicating low risk).

[0065] The application integrates the basic data and risk data to obtain corresponding feature data, then pre-processes the feature data to obtain corresponding processing data, then calls a pre-constructed risk assessment model, and then performs risk assessment processing on the processing data based on the risk assessment model to obtain a corresponding risk score. The application obtains feature data by integrating basic data and risk data, and obtains processing data by pre-processing the feature data, and then performs risk assessment processing on the processing data based on the use of the risk assessment model to obtain a corresponding risk score, so that the potential claim risk of the target enterprise can be efficiently and accurately predicted, and the accuracy of the obtained risk score is guaranteed. And subsequent generation of the premium of the target enterprise based on the obtained risk score is beneficial to improving the accuracy of the premium pricing.

[0066] In some optional implementations of the embodiment, before step S202, the electronic device can further perform the following steps:

[0067] The pre-constructed risk assessment model is called.

[0068] In the embodiment, the historical claim data can be obtained by collecting and organizing the claim data in a historical time period, including claim amount (the amount of each claim), claim reason (accident type, such as work injury, equipment failure, network attack, etc.), claim time (accident occurrence time, used for analyzing time trend), and collecting and organizing the enterprise data of each enterprise in the historical time period to obtain the historical enterprise data. The value of the historical time period is not specifically limited, and can be determined according to actual business needs, for example, it can be 2 years.

[0069] Based on the historical claim data and the historical enterprise data, corresponding sample data is constructed.

[0070] In the embodiment, the historical claim data and the historical enterprise feature data can be associated to form a complete data set, which serves as the sample data.

[0071] A preset light gradient boosting machine model is called.

[0072] In the embodiment, according to actual needs, the LGBM (LightGBM) model is selected as the basic model of the risk assessment model. The LGBM model can efficiently process high-dimensional data and nonlinear relationships, support direct input of category features, reduce coding workload, has the advantages of fast training speed, and is suitable for large-scale data.

[0073] A preset training strategy is obtained.

[0074] In the embodiment, the training strategy is to train the LGB model using historical data to learn the relationship between features and claim risk, and includes the following model training and optimization steps: 1) data division: divide the historical data into a training set (such as 70%) and a validation set (such as 30%) for model training and performance evaluation. 2) Hyperparameter adjustment: adjust key hyperparameters to optimize performance, for example: learning rate: control the contribution weight of each tree (such as 0.05). Tree depth: limit the maximum depth of a single tree (such as 6 layers). Leaf node number: control the complexity of the tree (such as 31 leaf nodes). Find the optimal parameter combination through grid search or random search. 3) Cross-validation: use k-fold cross-validation (such as k=5) to evaluate model stability and ensure generalization ability. Calculate the average performance indicators (such as AUC, RMSE) to verify the reliability of the model.

[0075] Based on the training strategy, the sample data is used to train and optimize the light gradient boosting machine model until a generated model that meets the construction target is obtained.

[0076] In the embodiment, the training and optimization of the model can be performed using the sample data according to the corresponding model training and optimization steps in the training strategy, so as to obtain a generated model meeting the requirements and serve as the required risk assessment model.

[0077] The generated model serves as the risk assessment model.

[0078] The application obtains historical claim data and historical enterprise data collected in advance, constructs corresponding sample data based on the historical claim data and the historical enterprise data, calls a preset light gradient boosting machine model, obtains a preset training strategy, trains and optimizes the light gradient boosting machine model based on the training strategy using the sample data until a generated model meeting the construction target is obtained, and then uses the generated model as the risk assessment model. The application can intelligently and accurately construct a risk assessment model meeting the construction target by constructing sample data based on the historical claim data and the historical enterprise data collected in advance and training and optimizing the light gradient boosting machine model based on the training strategy, improves the construction efficiency of the risk assessment model, and ensures the model effect of the obtained risk assessment model.

[0079] In some optional implementations, step S204 includes the following steps:

[0080] Obtain the industry code of the target industry and obtain the historical industry risk data of the target industry.

[0081] In the embodiment, the target industry can be queried for the industry code to obtain the matched industry code. In addition, the historical industry risk data of the target industry can be obtained from the public reports of industry associations, regulatory agencies or third-party risk databases. For example, the manufacturing industry can include equipment failure records, downtime length, maintenance costs, etc. The financial industry can include internal fraud cases, data leakage events, compliance violation records, etc.

[0082] The target generative adversarial network processes the industry code and the historical industry risk data to generate corresponding risk scenario data.

[0083] In the embodiment, the industry code and the historical industry risk data are input into the target generative adversarial network, and a large amount of simulated risk scenario data is generated by using the target generative adversarial network. For example, the manufacturing industry generator generates the following data: equipment failure type: mechanical failure (probability 60%), electrical failure (probability 40%). Downtime length: average 5 days, standard deviation 1 day. Maintenance cost: average 100,000 yuan, standard deviation 20,000 yuan.

[0084] extract a key risk indicator from the risk scenario data.

[0085] In the embodiment, the key risk indicators can be extracted from the risk scenario data generated by the target generative adversarial network, and the corresponding key risk indicators can be obtained, for example, the manufacturing industry: equipment failure frequency (times / year), average downtime (days), average maintenance cost (ten thousand yuan), the financial industry: internal fraud probability (times / year), average loss amount (ten thousand yuan), compliance violation times (times / year).

[0086] The key risk indicators are calculated based on a preset risk calculation strategy to obtain corresponding calculation results.

[0087] In the embodiment, the risk calculation strategy can specifically adopt a weighted summation strategy. All the extracted key risk indicators can be weighted and summed based on the risk calculation strategy, so as to obtain the corresponding calculation results as the industry risk coefficient. For example, the risk coefficient calculation formula of the manufacturing industry can include: Cz=w1×failure frequency+w2×average downtime+w3×average maintenance cost, wherein w1, w2, w3 are weights, reflecting the influence degree of different indicators on the risk. In addition, the industry risk coefficient can also be normalized to make its range between 0 and 1, which is convenient for subsequent premium calculation. For example, the manufacturing industry risk coefficient is normalized to 0.8 (high risk), and the financial industry risk coefficient is normalized to 0.6 (medium risk).

[0088] The calculation result is taken as the industry risk coefficient of the target industry.

[0089] In the embodiment, the industry risk coefficient calculation can convert the generated data of the generative adversarial network into quantifiable risk indicators, thereby providing a basis for premium pricing.

[0090] The application obtains the industry code of the target industry and the historical industry risk data of the target industry, then processes the industry code and the historical industry risk data based on the target generative adversarial network to generate corresponding risk scenario data, extracts key risk indicators from the risk scenario data, processes the key risk indicators based on a preset risk calculation strategy to obtain corresponding calculation results, and regards the calculation results as the industry risk coefficient of the target industry. The application obtains the industry code of the target industry and the historical industry risk data of the target industry, then processes the industry code and the historical industry risk data based on the target generative adversarial network to generate corresponding risk scenario data, extracts key risk indicators from the risk scenario data, and processes the key risk indicators based on a preset risk calculation strategy, so that the industry risk coefficient of the target industry can be quickly and accurately generated, the accuracy of the obtained industry risk coefficient is ensured, and the accuracy of the premium pricing can be improved by generating the premium of the target enterprise based on the obtained industry risk coefficient.

[0091] In some optional implementations, step S205 includes the following steps:

[0092] Obtaining the initial industry weight of the target industry and the risk scenario data corresponding to the industry risk coefficient.

[0093] In this embodiment, an initial industry weight corresponding to the industry historical risk level of the target industry can be preset. For example, the initial industry weight of the manufacturing industry is set to 0.3 because the equipment failure risk of the manufacturing industry is high, and the initial industry weight of the financial industry is set to 0.2 because the compliance risk of the financial industry is low. The risk scenario data is the risk scenario data corresponding to the target industry generated in the process of generating the industry risk coefficient of the target industry based on the target generative adversarial network.

[0094] Calling a preset dynamic weight optimization strategy.

[0095] In the present embodiment, the dynamic weight optimization strategy described above is a weight allocation strategy based on federated learning and Nash equilibrium. The dynamic weight optimization strategy combines federated learning and game theory to achieve the optimal solution of data privacy protection and premium pricing. Specifically, the strategy content of the dynamic weight optimization strategy includes: 1. Federated learning data training. Goal: Through the federated learning framework, use the local data of each enterprise to train the LGBM sub-model, while protecting the data privacy. Specific steps: 1) Local model training: Each enterprise uses local data to train an independent LGBM sub-model. During the training process, the enterprise only shares the model parameters (such as gradient, weight), not the original data, ensuring data privacy. For example, enterprise A trains sub-model MA using its historical claim data and feature data, and enterprise B trains sub-model MB. 2) Global parameter aggregation: The center server regularly collects the parameters (such as tree structure, leaf node weight) of each enterprise sub-model. Through federated averaging (FedAvg) or other aggregation algorithms, the sub-model parameters are weighted and averaged to update the global model. 3) Privacy protection: Use encryption technology (such as homomorphic encryption) or differential privacy technology to ensure data security during parameter transmission and aggregation. For example, by adding noise (differential privacy) or encrypting parameters (homomorphic encryption), the center server is prevented from inferring the original data of the enterprise. Among them, federated learning data training provides data basis for weight optimization, ensuring the global consistency of model parameters.

[0096] 2. Nash equilibrium solving. Goal: Solve optimal weight combination (a, b, g,) through game theory modeling, a, b are enterprise-specific weights, g, is industry adjustment weight, balance insurer and industry goals. Steps: 1) Game modeling. Insurer goal Uinsurer: Maximize profit: achieved through reasonable pricing (premium income) and risk control (claim payout). Control risk coverage: ensure premium income covers potential claim risk. Industry goal Uindustry: Balance industry overall risk: adjust premium through industry risk coefficient Cindustry. Premium fairness: ensure enterprises with different risk levels pay reasonable premiums. Model weight optimization problem as a multi-party game, where (a, b, g,) are game variables. 2) Equilibrium solving: Iterative optimization algorithm: use gradient descent or other optimization algorithms to iteratively adjust weights (a, b, g,) to approximate Nash equilibrium. For example, fix other weights and optimize a to maximize insurer goal Uinsurer; then fix a and optimize b to maximize industry goal Uindustry. Convergence condition: when weight adjustment is less than a preset threshold (e.g. 0.001), consider reaching Nash equilibrium. For example, if Da < 0.001 and Db < 0.001, stop iteration. 3) Optimal weight combination: output equilibrium solution (a*, b*, g,) as the optimal weight for premium pricing. For example, a* = 0.6, b* = 0.3, g, * = 0.1 (manufacturing). Nash equilibrium solving balances multiple goals through game theory methods, ensuring the fairness and reasonableness of weight combination, thereby avoiding cross-subsidization between industries (e.g. low-risk IT companies are not dragged down by high-risk manufacturing).

[0097] Based on the risk scenario data, the initial industry weights are optimized using the dynamic weight optimization strategy to obtain corresponding processing weights.

[0098] In this embodiment, the optimization of the initial industry weights can be based on the risk scenario data and the strategy content of the dynamic weight optimization strategy, and the optimal weight combination is solved, and the industry adjustment weight contained in the optimal weight combination, i.e. the processing weight, is extracted to be used as the final target industry adjustment weight. The dynamic weight optimization strategy enables the system to self-optimize based on real-time data, quickly adjusting the premium strategy when market environment, customer behavior or claim frequency changes.

[0099] The weights are used as the industry adjustment weights of the target industry.

[0100] In this embodiment, the adjustment of the industry adjustment weight introduces industry-specific risk, which is beneficial to ensure that the premium is consistent with the industry risk level.

[0101] The application obtains the initial industry weight of the target industry and the risk scenario data corresponding to the industry risk coefficient, calls a preset dynamic weight optimization strategy, optimizes the initial industry weight based on the risk scenario data and using the dynamic weight optimization strategy to obtain a corresponding processing weight, and takes the weight as the industry adjustment weight of the target industry. The application optimizes the initial industry weight based on the obtained initial industry weight of the target industry and the dynamic weight optimization strategy, thereby intelligently and accurately generating the industry adjustment weight of the target industry, ensuring the accuracy and rationality of the obtained industry adjustment weight, and facilitating the generation of the premium of the target enterprise based on the obtained industry adjustment weight, thereby improving the accuracy of premium pricing.

[0102] In some optional implementations of the embodiment, step S206 includes the following steps:

[0103] An initialized first enterprise weight corresponding to the target enterprise is obtained.

[0104] In the embodiment, the first enterprise weight includes a first initial weight (a) corresponding to a risk score and a second initial weight (b) corresponding to an industry risk coefficient, which are set in advance according to actual needs.

[0105] A preset reinforcement learning strategy is obtained.

[0106] In the embodiment, the reinforcement learning strategy is a strategy for dynamically adjusting the weight through reinforcement learning to reflect the enterprise-specific risk, which dynamically adjusts the enterprise weight according to the historical risk performance (such as the injury rate and the claim amount) of the enterprise and normalizes the adjusted enterprise weight, i.e., ensures a+b=1 to maintain the rationality of the weight combination. For example, if the injury rate of the enterprise is high (e.g., more than 20% of the industry average), the first initial weight is increased (e.g., a=0.7 is set), and the influence of the risk score of the risk assessment model is emphasized. If the claim amount of the enterprise fluctuates greatly (e.g., the standard deviation exceeds the industry average), the second initial weight is increased (e.g., b=0.6 is set), and the influence of the industry risk coefficient is emphasized.

[0107] The first enterprise weight is adjusted based on the reinforcement learning strategy to obtain a corresponding second enterprise weight.

[0108] In the embodiment, the first enterprise weight is adjusted according to the content of the reinforcement learning strategy to obtain the adjusted second enterprise weight, which is taken as the enterprise individual weight of the target enterprise to complete the optimization of the enterprise weight of the target enterprise.

[0109] The second enterprise weight is taken as the enterprise individuality weight of the target enterprise.

[0110] In the embodiment, the generated enterprise individuality weight captures the enterprise-specific risk of the target enterprise through reinforcement learning, thereby effectively improving the individualization of the premium.

[0111] The application obtains the initialized first enterprise weight corresponding to the target enterprise, then obtains the preset reinforcement learning strategy, then adjusts the first enterprise weight based on the reinforcement learning strategy to obtain the corresponding second enterprise weight, and then takes the second enterprise weight as the enterprise individuality weight of the target enterprise. The application obtains the initialized first enterprise weight corresponding to the target enterprise, and then adjusts the first enterprise weight based on the use of the reinforcement learning strategy and takes the obtained adjusted second enterprise weight as the enterprise individuality weight of the target enterprise, so that the optimization of the first enterprise weight can be accurately and intelligently completed, and the enterprise individuality weight of the target enterprise is generated, thereby effectively ensuring the rationality and accuracy of the generated enterprise individuality weight, which is beneficial to generating the premium of the target enterprise based on the obtained enterprise individuality weight, and thereby improving the accuracy of the premium pricing.

[0112] In some optional implementation manners of the embodiment, before step S207, the above-mentioned electronic device can further perform the following steps:

[0113] Obtain the enterprise scale information of the target enterprise.

[0114] In the embodiment, the corresponding scale level can be divided according to the number of employees of the target enterprise to obtain the enterprise scale information. For example, small enterprise <100 people, medium enterprise 100-500 people, large enterprise >500 people. The more the number of employees, the higher the benchmark premium (for example, the benchmark premium of large enterprise is 2 times that of small enterprise).

[0115] Obtain the regional information of the target enterprise.

[0116] In the embodiment, the corresponding regional level can be divided according to the region of the target enterprise to obtain the regional information. For example, first-tier city, second-tier city, third-tier city. And the benchmark premium of the first-tier city is higher, for example, the benchmark premium of the first-tier city is 1.5 times that of the third-tier city.

[0117] Call the preset benchmark premium calculation strategy.

[0118] In the embodiment, the strategy content of the above-mentioned base premium calculation strategy includes: the base premium Base of the enterprise is a weighted combination of the scale and the region of the enterprise, that is, Base = scale coefficient * region coefficient * base premium. For example, the base premium of a large enterprise (scale coefficient = 2) in a first-tier city (region coefficient = 1.5) is: Base = 2 * 1.5 * 1000 = 3000.

[0119] The enterprise scale information and the region information are calculated based on the base premium calculation strategy to obtain the base premium of the target enterprise.

[0120] In the embodiment, the enterprise scale information can be mapped to the corresponding enterprise scale coefficient and the region information can be mapped to the corresponding region coefficient according to the actual mapping relationship. Then, the enterprise scale coefficient and the region coefficient obtained by mapping are calculated based on the strategy content of the above-mentioned base premium calculation strategy, and the calculation result is taken as the base premium of the target enterprise. The generated base premium is used to provide a basis for premium pricing and reflects the general risk of the scale and region of the target enterprise.

[0121] The enterprise scale information of the target enterprise is obtained, and the region information of the target enterprise is obtained. Then, a preset base premium calculation strategy is called. The enterprise scale information and the region information are calculated based on the base premium calculation strategy to obtain the base premium of the target enterprise. The enterprise scale information and the region information are calculated based on the use of the base premium calculation strategy, so that the base premium of the target enterprise can be automatically and accurately generated, and the accuracy of the obtained base premium is ensured. This is conducive to generating the premium of the target enterprise based on the obtained base premium, and the accuracy of the premium pricing can be improved.

[0122] In some optional implementations, the obtained user information is subject to user consent and complies with relevant laws and relevant policies.

[0123] In addition, the non-company software tools or components appearing in the embodiments of the present application are only examples and do not represent actual use.

[0124] It should be understood that the size of the serial number of each step in the above-mentioned embodiments does not mean the execution order, and the execution order of each process should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0125] It should be emphasized that, in order to further ensure the privacy and security of the above-mentioned target premium, the above-mentioned target premium can also be stored in a node of a block chain.

[0126] The blockchain referred to in the present application is a new application mode of distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm and other computer technologies. The blockchain is essentially a decentralized database, which is a series of data blocks associated using cryptographic methods, each data block containing information of a batch of network transactions, used to verify the validity of the information (anti-fake) and generate the next block. The blockchain can include a blockchain underlying platform, a platform product service layer, and an application service layer.

[0127] The embodiments of the present application can acquire and process related data based on artificial intelligence technology. Artificial intelligence (AI) is the use of digital computers or computer-controlled machines to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0128] Artificial intelligence basic technologies generally include technologies such as sensors, special artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction system, mechatronics, etc. Artificial intelligence software technology mainly includes computer vision technology, robot technology, biometric technology, speech processing technology, natural language processing technology, and machine learning / deep learning, etc.

[0129] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by computer readable instructions instructing related hardware, and the computer readable instructions can be stored in a computer readable storage medium. The program can include the processes of the above-mentioned embodiments when executed, wherein the storage medium can be a non-volatile storage medium such as a magnetic disc, an optical disc, a read-only memory (ROM), or a random access memory (RAM).

[0130] It should be understood that although each step in the flowchart of the accompanying drawings is displayed in sequence according to the arrow, these steps are not necessarily executed in sequence according to the arrow. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and they can be executed in other orders. Moreover, at least part of the steps in the flowchart of the accompanying drawings can include multiple sub-steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence is not necessarily sequential, but can be executed in rotation or alternation with at least part of other steps or sub-steps or stages of other steps.

[0131] Further reference Figure 3 , as the implementation of the method shown in the above Figure 2 , the present application provides an embodiment of a data generation device based on artificial intelligence, which device embodiment corresponds to the method embodiment shown in Figure 2 , the device can be applied to various electronic devices.

[0132] As shown in Figure 3 , the data generation device based on artificial intelligence 300 of the embodiment comprises a first acquisition module 301, a first generation module 302, a first processing module 303, a second generation module 304, a third generation module 305, a fourth generation module 306 and a second processing module 307. Among them:

[0133] The first acquisition module 301 is used for acquiring enterprise data of a target enterprise;

[0134] The first generation module 302 is used for performing risk assessment processing on the enterprise data based on a preset risk assessment model, and generating a corresponding risk score;

[0135] The first processing module 303 is used for determining a target industry corresponding to the target enterprise, and calling a target generative adversarial network corresponding to the target industry;

[0136] The second generation module 304 is used for generating an industry risk coefficient corresponding to the target industry based on the target generative adversarial network;

[0137] The third generation module 305 is used for generating an industry adjustment weight corresponding to the target industry based on the industry risk coefficient;

[0138] The fourth generation module 306 is used for generating an enterprise individual weight corresponding to the target enterprise based on a preset weight generation strategy;

[0139] The second processing module 307 is used for calculating the base premium of the target enterprise, the risk score, the industry adjustment weight and the enterprise individual weight based on a preset premium calculation formula, and obtaining a target premium of the target enterprise.

[0140] In the embodiment, the operations performed by the above-mentioned modules or units correspond one by one to the steps of the data generation method based on artificial intelligence of the foregoing embodiments, which will not be repeated here.

[0141] In some optional implementations of the embodiment, the enterprise data comprises basic data and risk data; the first generation module 302 comprises:

[0142] The integration sub-module is used for integrating the basic data and the risk data to obtain corresponding feature data;

[0143] a preprocessing submodule, configured to preprocess the feature data to obtain corresponding processing data;

[0144] a first calling submodule, configured to call a pre-constructed risk assessment model;

[0145] an assessment submodule, configured to perform risk assessment processing on the processing data based on the risk assessment model to obtain a corresponding risk score.

[0146] In the embodiment, the modules or units are respectively used for performing operations corresponding to the steps of the data generation method based on artificial intelligence of the foregoing embodiments, and thus will not be described here again.

[0147] In some optional implementations of the embodiment, the data generation apparatus based on artificial intelligence further includes:

[0148] a second obtaining module, configured to obtain pre-collected historical claim data and historical enterprise data;

[0149] a construction module, configured to construct corresponding sample data based on the historical claim data and the historical enterprise data;

[0150] a first calling module, configured to call a pre-set light gradient boosting machine model;

[0151] a third obtaining module, configured to obtain a pre-set training strategy;

[0152] a training module, configured to train and optimize the light gradient boosting machine model based on the training strategy and using the sample data until a generation model meeting a construction target is obtained;

[0153] a determination module, configured to determine the generation model as the risk assessment model.

[0154] In the embodiment, the modules or units are respectively used for performing operations corresponding to the steps of the data generation method based on artificial intelligence of the foregoing embodiments, and thus will not be described here again.

[0155] In some optional implementations of the embodiment, the second generation module 304 includes:

[0156] a second obtaining submodule, configured to obtain an industry code of the target industry and obtain historical industry risk data of the target industry;

[0157] a processing submodule, configured to process the industry code and the historical industry risk data based on the target generative adversarial network to generate corresponding risk scenario data;

[0158] extracting a key risk indicator from the risk scenario data;

[0159] computing the key risk indicator based on a preset risk computing strategy to obtain a corresponding computing result;

[0160] a first determining sub-module configured to take the computing result as an industry risk coefficient of the target industry.

[0161] In the embodiment, the modules or units are respectively used for operations corresponding to the steps of the data generation method based on artificial intelligence in the foregoing embodiments, and thus will not be described herein.

[0162] In some optional implementations of the embodiment, the third generating module 305 includes:

[0163] a third acquiring sub-module configured to acquire an initial industry weight of the target industry and acquire the risk scenario data corresponding to the industry risk coefficient;

[0164] a second calling sub-module configured to call a preset dynamic weight optimization strategy;

[0165] an optimization sub-module configured to perform optimization processing on the initial industry weight based on the risk scenario data and using the dynamic weight optimization strategy to obtain a corresponding processing weight;

[0166] a second determining sub-module configured to take the weight as an industry adjustment weight of the target industry.

[0167] In the embodiment, the modules or units are respectively used for operations corresponding to the steps of the data generation method based on artificial intelligence in the foregoing embodiments, and thus will not be described herein.

[0168] In some optional implementations of the embodiment, the fourth generating module 306 includes:

[0169] a fourth acquiring sub-module configured to acquire an initialized first enterprise weight corresponding to the target enterprise;

[0170] a fifth acquiring sub-module configured to acquire a preset reinforcement learning strategy;

[0171] an adjusting sub-module configured to perform adjustment processing on the first enterprise weight based on the reinforcement learning strategy to obtain a corresponding second enterprise weight;

[0172] a third determining sub-module configured to take the second enterprise weight as an enterprise individual weight of the target enterprise.

[0173] In the embodiment, the operations performed by the modules or units described above correspond to the steps of the artificial intelligence-based data generation method of the foregoing embodiments one by one, and will not be described here again.

[0174] In some optional implementations of the embodiment, the artificial intelligence-based data generation apparatus further includes:

[0175] A fourth acquisition module is configured to acquire enterprise scale information of the target enterprise.

[0176] A fifth acquisition module is configured to acquire regional information of the target enterprise.

[0177] A second calling module is configured to call a preset benchmark premium calculation strategy.

[0178] A calculation module is configured to perform calculation processing on the enterprise scale information and the regional information based on the benchmark premium calculation strategy, to obtain a benchmark premium of the target enterprise.

[0179] In the embodiment, the operations performed by the modules or units described above correspond to the steps of the artificial intelligence-based data generation method of the foregoing embodiments one by one, and will not be described here again.

[0180] To solve the above technical problems, the embodiment of the present application further provides a computer device. For details, please refer to Figure 4 , Figure 4 The basic structure block diagram of the computer device of the embodiment is shown in the figure.

[0181] The computer device 4 includes a memory 41, a processor 42, and a network interface 43, which are connected to each other through a system bus. It should be pointed out that only the computer device 4 with components 41-43 is shown in the figure, but it should be understood that it is not required to implement all the shown components, and more or fewer components can be alternatively implemented. Among them, those skilled in the art can understand that the computer device here is a device capable of automatically performing numerical calculation and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), digital signal processors (DSPs), embedded devices, etc.

[0182] The computer device can be a desktop computer, a notebook computer, a palm computer, and a cloud server, etc. The computer device can interact with the user through a keyboard, a mouse, a remote controller, a touchpad, a voice control device, etc.

[0183] The memory 41 includes at least one type of readable storage medium, such as a flash memory, a hard disk, a multimedia card, a card-type memory (e.g., an SD or DX memory, etc.), a random access memory (RAM), a static random access memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, an optical disk, etc. In some embodiments, the memory 41 can be an internal storage unit of the computer device 4, such as a hard disk or a memory of the computer device 4. In other embodiments, the memory 41 can also be an external storage device of the computer device 4, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the computer device 4. Of course, the memory 41 can also include both the internal storage unit and the external storage device of the computer device 4. In this embodiment, the memory 41 is generally used to store an operating system and various application software installed on the computer device 4, such as computer readable instructions of the artificial intelligence-based data generation method, etc. In addition, the memory 41 can also be used to temporarily store various data that have been output or will be output.

[0184] The processor 42 can be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip in some embodiments. The processor 42 is generally used to control the overall operation of the computer device 4. In this embodiment, the processor 42 is used to run computer readable instructions or process data stored in the memory 41, such as computer readable instructions of the artificial intelligence-based data generation method.

[0185] The network interface 43 can include a wireless network interface or a wired network interface, and is generally used to establish a communication connection between the computer device 4 and other electronic devices.

[0186] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0187] In the embodiments of the present application, after obtaining the enterprise data of the target enterprise, the risk score is generated by performing risk assessment processing on the enterprise data based on the use of the risk assessment model, then the industry risk coefficient corresponding to the target industry is generated by using the target generation adversarial network corresponding to the target industry, and the industry adjustment weight of the target industry is generated based on the use of the industry risk coefficient, and then the enterprise individual weight of the target enterprise is generated based on the use of the weight generation strategy, and finally the base premium, the risk score, the industry adjustment weight and the enterprise individual weight of the target enterprise are calculated and processed based on the use of the premium calculation formula, so that the target premium of the target enterprise can be automatically and accurately generated. The present application can realize accurate, fair and reasonable premium pricing by combining the multi-level factors corresponding to the base premium, the risk score, the industry adjustment weight and the enterprise individual weight of the target enterprise for the calculation and processing of the premium, and effectively improves the accuracy of the generated target premium.

[0188] The present application also provides another embodiment, that is, a computer readable storage medium storing computer readable instructions executable by at least one processor to cause the at least one processor to perform the steps of the artificial intelligence-based data generation method as described above.

[0189] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0190] In the embodiments of the present application, after obtaining the enterprise data of the target enterprise, the risk score is generated by performing risk assessment processing on the enterprise data based on the use of the risk assessment model, then the industry risk coefficient corresponding to the target industry is generated by using the target generation adversarial network corresponding to the target industry, and the industry adjustment weight of the target industry is generated based on the use of the industry risk coefficient, and then the enterprise individual weight of the target enterprise is generated based on the use of the weight generation strategy, and finally the base premium, the risk score, the industry adjustment weight and the enterprise individual weight of the target enterprise are calculated and processed based on the use of the premium calculation formula, so that the target premium of the target enterprise can be automatically and accurately generated. The present application can realize accurate, fair and reasonable premium pricing by combining the multi-level factors corresponding to the base premium, the risk score, the industry adjustment weight and the enterprise individual weight of the target enterprise for the calculation and processing of the premium, and effectively improves the accuracy of the generated target premium.

[0191] Those skilled in the art can clearly understand the above-mentioned embodiment method can be realized by means of software and the necessary general hardware platform, of course, it can also be realized by hardware, but in many cases, the former is a better embodiment. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes a plurality of instructions for making a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) execute the methods described in various embodiments of the present application.

[0192] Obviously, the above-described embodiments are only some of the embodiments of the present application, not all the embodiments, and the drawings show the preferred embodiments of the present application, but do not limit the patent scope of the present application. The present application can be implemented in many different forms, and conversely, the purpose of providing these embodiments is to make the disclosure of the present application more thorough and comprehensive. Although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions recorded in the foregoing specific embodiments, or make equivalent replacements to some of the technical features. Any equivalent structure made by using the contents of the specification and drawings, directly or indirectly applied to other related technical fields, is also within the scope of the patent protection of the present application.

Claims

1. A data generation method based on artificial intelligence, characterized in that: The steps include: Obtain enterprise data of target enterprises; Perform risk assessment on the enterprise data based on a preset risk assessment model to generate a corresponding risk score; Determine a target industry corresponding to the target enterprise, and call a target generative adversarial network corresponding to the target industry; Generate an industry risk coefficient corresponding to the target industry based on the target generative adversarial network; generating an industry adjustment weight corresponding to the target industry based on the industry risk coefficient; Generate an enterprise personality weight corresponding to the target enterprise based on a preset weight generation strategy; The target premium of the target enterprise is obtained by calculating the base premium of the target enterprise, the risk score, the industry adjustment weight and the enterprise individual weight based on a preset premium calculation formula.

2. The artificial intelligence-based data generation method according to claim 1, characterized in that: The enterprise data includes basic data and risk data; The step of performing risk assessment processing on the enterprise data based on a preset risk assessment model to generate a corresponding risk score specifically includes: Integrate the basic data and risk data to obtain corresponding feature data; Preprocessing the characteristic data to obtain corresponding processed data; Calling pre-built risk assessment models; Perform risk assessment on the processed data based on the risk assessment model to obtain a corresponding risk score.

3. The artificial intelligence-based data generation method according to claim 1, characterized in that: The enterprise data includes basic data and risk data; Before the step of performing risk assessment on the enterprise data based on a preset risk assessment model to generate a corresponding risk score, the method further includes: Obtain pre-collected historical claims data and historical enterprise data; Constructing corresponding sample data based on the historical claims data and the historical enterprise data; Call the preset lightweight gradient boosting machine model; Get the preset training strategy; Based on the training strategy, the lightweight gradient boosting machine model is trained and optimized using the sample data until a generation model that meets the construction goal is obtained; The generated model is used as the risk assessment model.

4. The artificial intelligence-based data generation method according to claim 1, characterized in that: The step of generating an industry risk coefficient corresponding to the target industry based on the target generative adversarial network specifically includes: Obtaining the industry code of the target industry and obtaining historical industry risk data of the target industry; Processing the industry code and the historical industry risk data based on the target generative adversarial network to generate corresponding risk scenario data; extracting key risk indicators from the risk scenario data; Calculate the key risk indicators based on a preset risk calculation strategy to obtain corresponding calculation results; The calculation result is used as the industry risk coefficient of the target industry.

5. The artificial intelligence-based data generation method according to claim 4, characterized in that: The step of generating the industry adjustment weight corresponding to the target industry based on the industry risk coefficient specifically includes: Obtaining the initial industry weight of the target industry, and obtaining the risk scenario data corresponding to the industry risk coefficient; Call the preset dynamic weight optimization strategy; Based on the risk scenario data, the initial industry weights are optimized using the dynamic weight optimization strategy to obtain corresponding processing weights; The weight is used as the industry adjustment weight of the target industry.

6. The artificial intelligence-based data generation method according to claim 1, characterized in that: The step of generating the enterprise personality weight corresponding to the target enterprise based on the preset weight generation strategy specifically includes: Obtaining an initialized first enterprise weight corresponding to the target enterprise; Get the preset reinforcement learning strategy; Adjusting the first enterprise weight based on the reinforcement learning strategy to obtain a corresponding second enterprise weight; The second enterprise weight is used as the enterprise personality weight of the target enterprise.

7. The artificial intelligence-based data generation method according to claim 1, characterized in that: Before the step of calculating the target enterprise's base premium, the risk score, the industry adjustment weight, and the enterprise individual weight based on a preset premium calculation formula to obtain the target premium of the target enterprise, the method further includes: Obtaining enterprise scale information of the target enterprise; Obtaining the geographical information of the target enterprise; Call the preset benchmark premium calculation strategy; The enterprise size information and the regional information are calculated and processed based on the benchmark premium calculation strategy to obtain a benchmark premium for the target enterprise.

8. A data generation device based on artificial intelligence, characterized in that: include: A first acquisition module is used to acquire enterprise data of a target enterprise; A first generating module is used to perform risk assessment processing on the enterprise data based on a preset risk assessment model to generate a corresponding risk score; A first processing module is configured to determine a target industry corresponding to the target enterprise and invoke a target generative adversarial network corresponding to the target industry; A second generating module is configured to generate an industry risk coefficient corresponding to the target industry based on the target generative adversarial network; A third generating module is configured to generate an industry adjustment weight corresponding to the target industry based on the industry risk coefficient; A fourth generating module, configured to generate an enterprise personality weight corresponding to the target enterprise based on a preset weight generating strategy; The second processing module is used to calculate the benchmark premium of the target enterprise, the risk score, the industry adjustment weight and the enterprise individual weight based on a preset premium calculation formula to obtain the target premium of the target enterprise.

9. A computer device, characterized in that: The method comprises a memory and a processor, wherein the memory stores computer-readable instructions, and when the processor executes the computer-readable instructions, the steps of the artificial intelligence-based data generation method as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of the artificial intelligence-based data generation method according to any one of claims 1 to 7.

Images