Risk assessment method and device, computer equipment and storage medium

Abstract

The embodiment of the invention belongs to the field of artificial intelligence, and relates to a risk assessment method, which comprises the steps of obtaining a transaction historical data set which is composed of insurance policy information and user information, performing retention prediction processing on the user information to obtain a user retention rate, and inputting the user retention rate and the insurance policy information into a preset regression prediction model for calculation to obtain a customer life cycle value, and performing simulation sampling analysis on the customer life cycle value to obtain user risk assessment. The invention further provides a risk assessment device, computer equipment and a storage medium. The technical problem that traditional risk assessment only depends on the current insurance policy value of the customer and neglects the long-term value is solved, so that an insurance company can more accurately identify and distinguish high-risk and low-risk customers, and optimal allocation of resources and accurate management of risks are further realized.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to risk assessment methods, devices, computer equipment and storage media. Background Art

[0002] Customer value assessment is a crucial link in insurance marketing. As an important measurement standard in refined operations, customer value assessment can help companies reasonably allocate differentiated strategy deployment under limited resources to maximize company benefits.

[0003] Determining customer delivery strategies based on the risk level of a single product during customer value assessment will not only result in the loss of high-quality customers and affect the quality of existing customers, but will also fail to create profits for the company in the long run and maximize the investment-output ratio.

[0004] However, when assessing customer risks, the traditional insurance industry often only considers the customer's current value, that is, whether the customer's current policy is profitable or not, but is unable to make an overall assessment of the value that the customer can bring to the company throughout its life cycle. Summary of the invention

[0005] The purpose of the embodiments of the present application is to propose a risk assessment method, apparatus, computer equipment and storage medium to solve the technical problem that the traditional insurance industry is unable to assess the entire life cycle of the customer in an overall risk assessment manner when assessing customer risks.

[0006] In order to solve the above technical problems, the present application embodiment provides a risk assessment method, which adopts the following technical solution:

[0007] Acquire a transaction history data set, wherein the transaction history data set consists of policy information and user information;

[0008] Performing retention prediction processing on the user information to obtain a user retention rate;

[0009] Input the user retention rate and the policy information into a preset regression prediction model for calculation to obtain the customer life cycle value;

[0010] A simulated sampling analysis is performed on the customer lifetime value to obtain a user risk assessment.

[0011] Furthermore, the step of performing retention prediction processing on the user information to obtain the user retention rate specifically includes:

[0012] Performing factor processing on the user information to obtain a user transaction factor;

[0013] The user transaction factor is input into a preset LGBM model for prediction processing to obtain the user retention rate.

[0014] Furthermore, before the step of inputting the user retention rate and the policy information into a preset regression prediction model for calculation to obtain the customer lifetime value, the method further includes:

[0015] Obtaining the acquisition time node of the user information and the regression prediction model to be trained;

[0016] Pull out the user annual information of the year before the acquisition time node from the historical transaction database as a user training set;

[0017] Extracting time nodes from the user's annual information to obtain annual time features;

[0018] Estimating the annual time characteristics to obtain model training parameters;

[0019] The model training parameters are imported into the regression prediction model to be trained, and the regression prediction model to be trained is trained using the user training set to obtain a preset regression prediction model.

[0020] Furthermore, after the step of importing the model training parameters into the regression prediction model to be trained, and training the regression prediction model to be trained using the user training set to obtain a preset regression prediction model, the method further includes:

[0021] Obtaining a user test set from a historical transaction database, importing the user test set into the preset regression prediction model, and outputting a prediction result;

[0022] Based on the prediction results and the preset standard results, the back propagation algorithm is used for fitting to obtain the prediction error;

[0023] The prediction error is compared with a preset threshold. If the prediction error is greater than the preset threshold, the preset regression prediction model is iteratively updated until the prediction error is less than or equal to the preset threshold, and the preset regression prediction model is updated.

[0024] Furthermore, the preset regression prediction model includes an input layer, a processing layer and an output layer, and the step of inputting the user retention rate and the policy information into the preset regression prediction model for calculation to obtain the customer lifetime value specifically includes:

[0025] Extracting features of the insurance policy information through the input layer to obtain an insurance policy feature tensor;

[0026] Performing a life cycle value calculation on the policy feature tensor and the user retention rate through the processing layer to obtain the customer life cycle value;

[0027] The customer lifetime value is outputted through the output layer.

[0028] Furthermore, the step of performing simulated sampling analysis on the customer lifecycle value to obtain a user risk assessment specifically includes:

[0029] Randomly sampling the customer life cycle value to obtain a user value path;

[0030] The user value path is processed for predictive distribution through Bayesian inference method to obtain the user risk assessment.

[0031] Furthermore, after the step of performing a simulated sampling analysis on the customer life cycle value to obtain a user risk assessment, the method further includes:

[0032] Based on the user risk assessment, the user groups are divided into value groups to obtain target user value groups;

[0033] A differentiated strategy is deployed for the user value groups to obtain a deployment strategy.

[0034] In order to solve the above technical problems, the embodiment of the present application further provides a risk assessment device, which adopts the following technical solution:

[0035] An acquisition module is used to acquire a transaction history data set, wherein the transaction history data set consists of policy information and user information;

[0036] A processing module, used to perform retention prediction processing on the user information to obtain a user retention rate;

[0037] A calculation module, used for inputting the user retention rate and the policy information into a preset regression prediction model for calculation to obtain the customer life cycle value;

[0038] The analysis module is used to perform simulated sampling analysis on the customer life cycle value to obtain a user risk assessment.

[0039] In order to solve the above technical problems, the embodiment of the present application further provides a computer device, which adopts the following technical solution:

[0040] A computer device comprises a memory and a processor, wherein the memory stores computer-readable instructions, and the processor implements the steps of the risk assessment method when executing the computer-readable instructions.

[0041] In order to solve the above technical problems, the embodiment of the present application further provides a computer-readable storage medium, which adopts the following technical solution:

[0042] A computer-readable storage medium, characterized in that computer-readable instructions are stored on the computer-readable storage medium, and when the computer-readable instructions are executed by a processor, the steps of the risk assessment method are implemented.

[0043] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0044] The embodiment of the present application obtains a transaction history data set consisting of policy information and user information, performs retention prediction processing on the user information to obtain a user retention rate, inputs the user retention rate and the policy information into a preset regression prediction model for calculation to obtain a customer lifetime value, performs simulated sampling analysis on the customer lifetime value to obtain a user risk assessment, calculates the customer lifetime value through retention prediction and regression prediction models, and performs risk assessment through simulated sampling analysis, so that insurance companies can more accurately identify and distinguish high-risk and low-risk customers, thereby achieving optimal allocation of resources and precise management of risks, allowing insurance companies to optimize customer structure and formulate more targeted customer maintenance and market strategies, thereby maximizing the benefits of insurance companies with limited resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0046] Figure 1 is an exemplary system architecture diagram to which the present application may be applied;

[0047] Figure 2 A flow chart of an embodiment of a risk assessment method according to the present application;

[0048] Figure 3 is a schematic structural diagram of an embodiment of a risk assessment device according to the present application;

[0049] Figure 4 It is a structural diagram of an embodiment of a computer device according to the present application. DETAILED DESCRIPTION

[0050] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0051] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0052] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0053] like Figure 1 As shown, the system architecture 100 may include a terminal device 101, a network 102 and a server 103. The terminal device 101 may be a laptop 1011, a tablet computer 1012 or a mobile phone 1013. The network 102 is used to provide a medium for a communication link between the terminal device 101 and the server 103. The network 102 may include various connection types, such as wired, wireless communication links or optical fiber cables.

[0054] The 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.

[0055] The terminal device 101 can be any electronic device with a display screen and supporting web browsing. In addition to a laptop computer 1011, a tablet computer 1012 or a mobile phone 1013, the terminal device 101 can also be an e-book reader, an MP3 player (Moving Picture Experts Group Audio Layer III), an MP4 (Moving Picture Experts Group Audio Layer IV), a laptop computer, a desktop computer, etc.

[0056] The server 103 may be a server that provides various services, such as a background server that provides support for web pages displayed on the terminal device 101 .

[0057] It should be noted that the risk assessment method provided in the embodiment of the present application is generally executed by a server / terminal device, and accordingly, the risk assessment device is generally set in the server / terminal device.

[0058] 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.

[0059] Continue to refer Figure 2 , showing a flow chart of an embodiment of a method for risk assessment according to the present application. The risk assessment method comprises the following steps:

[0060] Step S201, obtaining a transaction history data set.

[0061] In this embodiment, the risk assessment method can be deployed in a risk assessment platform. The risk assessment platform can be constructed by a server or a server cluster. The server or server cluster can be any electronic device with functions such as text recognition, text processing, data transmission, and data storage. The risk assessment method runs on the electronic device (for example, Figure 1 The server / terminal device shown in the figure can receive the transaction history data set through a wired connection or a wireless connection. It should be noted that the above wireless connection method may include but is not limited to 3G / 4G / 5G connection, WiFi connection, Bluetooth connection, WiMAX connection, Zigbee connection, UWB (ultra wideband) connection, and other wireless connection methods currently known or to be developed in the future.

[0062] In this embodiment, the transaction history data set may be a data set including policy information and user information. Specifically, the user information may include but is not limited to basic information such as the user's age, gender, region, wealth level, etc., customer stickiness information such as the user's product holdings, platform activity, intimacy with agents, and other dimensional information such as customer credit level and customer online transaction preferences; the policy information may include but is not limited to insurance information such as product type, policy status, effective premium, maturity payment, actual expenses, etc.

[0063] Step S202: Perform retention prediction processing on the user information to obtain the user retention rate.

[0064] In this embodiment, the user retention rate may be the probability that the user will continue to hold the insurance product or maintain the relationship with the insurance company at a certain point in time in the future. The user retention rate may be used to help insurance companies predict the long-term stability and profit potential of their customers.

[0065] Specifically, regarding the specific implementation process of performing retention prediction processing on user information to obtain the user retention rate, this application will provide further detailed descriptions of this in subsequent specific embodiments, and will not be elaborated on here.

[0066] Step S203: Input the user retention rate and the policy information into a preset regression prediction model for calculation to obtain the customer lifetime value.

[0067] In this embodiment, the above-mentioned preset regression prediction model can be a prediction model for predicting customer lifetime value (CLV) based on a framework built on machine learning models such as XGBoost (Extreme Gradient Boosting), Random Forest or ARIMA (Autoregressive Integrated Moving Average).

[0068] Specifically, the above customer lifetime value can be an indicator to measure the total contribution or total profit that a customer brings to the company during the entire period of cooperation. The calculation of the above customer lifetime value takes into account all cash flows from the first purchase to the end of the customer relationship, discounted to the current value. The above customer lifetime value reflects the long-term value of a customer to the company and helps companies identify which customer groups are most profitable, so as to carry out targeted marketing, service optimization and resource allocation.

[0069] In one possible embodiment, the customer lifetime value may be the total profit that the customer is expected to bring to the insurance company during the duration of the relationship between the customer and the insurance company. The customer lifetime value is based on a one-year cycle, taking into account the user's contribution to the policy within one year and the customer's retention expectations. The customer lifetime value may reflect the overall economic value of a customer from the initial joining to the loss or retention period.

[0070] Specifically, regarding the specific implementation process of inputting the user retention rate and policy information into the preset regression prediction model for calculation to obtain the customer lifetime value, this application will provide further detailed descriptions of this in subsequent specific embodiments and will not be elaborated in detail here.

[0071] Step S204, performing simulated sampling analysis on the customer life cycle value to obtain a user risk assessment.

[0072] In this embodiment, the above-mentioned user risk assessment may include assessment considerations of the uncertainty of the customer's future transaction behavior, possible credit risk, probability of policy renewal, and possibility of customer churn.

[0073] In one possible embodiment, by simulating different economic scenarios and customer behaviors, insurance companies can obtain the characteristics of high-risk and low-risk customers (i.e., the above-mentioned user risk assessment), and then adopt corresponding strategies to optimize resource allocation and reduce potential negative impacts, helping companies to form more accurate customer maintenance strategies, thereby maintaining a stable customer base and improving the company's overall profitability.

[0074] Specifically, regarding the above-mentioned simulated sampling analysis of customer lifetime value to obtain the specific implementation process of user risk assessment, this application will provide further detailed descriptions of this in subsequent specific embodiments, and will not be elaborated in detail here.

[0075] The present application obtains a transaction history data set consisting of policy information and user information, performs retention prediction processing on the user information to obtain a user retention rate, inputs the user retention rate and the policy information into a preset regression prediction model for calculation to obtain a customer lifetime value, performs simulated sampling analysis on the customer lifetime value to obtain a user risk assessment, calculates the customer lifetime value through retention prediction and regression prediction models, and performs risk assessment through simulated sampling analysis, so that insurance companies can more accurately identify and distinguish high-risk and low-risk customers, thereby achieving optimal allocation of resources and precise management of risks, allowing insurance companies to optimize customer structure and formulate more targeted customer maintenance and market strategies, thereby maximizing the benefits of insurance companies with limited resources.

[0076] In some optional implementations, step S202 includes the following steps:

[0077] Perform factor processing on user information to obtain user transaction factors;

[0078] The user transaction factor is input into the preset LGBM model for prediction processing to obtain the user retention rate.

[0079] In this embodiment, the above-mentioned user transaction factors may include but are not limited to the user's age, gender, region, wealth level, product holding type, policy status, effective premium, maturity payment, actual expenses incurred, purchase frequency, historical transaction behavior and other information.

[0080] In this embodiment, the above-mentioned preset LGBM model can be an efficient machine learning model based on the gradient boosting machine. The above-mentioned specially preset LGBM model is provided with preset prediction rules for processing classification or regression problems. The above-mentioned LGBM model can use the tree algorithm as its basic learner to improve the accuracy and robustness of the prediction by constructing multiple decision trees and integrating the prediction results.

[0081] In a possible embodiment, basic information and transaction history data of users, including age, gender, region, product type, policy status, effective premium and other data, can be collected from a historical transaction database, and the above data can be cleaned, and missing values ​​and outliers can be processed to ensure data quality; the above factor processing can include the following processing methods: for numerical features such as premium amount and compensation amount, standardization or normalization is performed; for categorical features such as gender and policy status, one-hot encoding (One-Hot Encoding) or label encoding (Label Encoding) is used to convert them into a format that can be processed by the machine learning model; for derived features, new features are created based on existing data, such as extracting the month or year from the transaction date, or calculating the customer's policy renewal rate, etc.; using statistical methods (such as correlation coefficient analysis) or model-based feature selection methods (such as using random forest feature importance scoring) to select the most influential features for predicting user retention rate from the above user transaction factors as input to the above preset LGBM model, and the preset LGBM model is used for prediction processing to obtain the above user retention rate.

[0082] This application can achieve high-precision user retention rate prediction by factor processing of user information and inputting the obtained user transaction factors into the preset LGBM model for prediction processing. By fine-tuning the user's historical behavior and transaction data, the key factors affecting user retention can be effectively captured, thereby making the prediction results more accurate and reliable.

[0083] In some optional implementations, before step S203, the following steps are further included:

[0084] Obtain the time node of obtaining user information and the regression prediction model to be trained;

[0085] Pull out the user annual information of the year before the time node from the historical transaction database as the user training set;

[0086] Extract time nodes from user annual information to obtain annual time features;

[0087] Estimate annual time characteristics and obtain model training parameters;

[0088] The model training parameters are imported into the regression prediction model to be trained, and the regression prediction model to be trained is trained using the user training set to obtain a preset regression prediction model.

[0089] In this embodiment, the above-mentioned user annual information may be the user information of the year before the current acquisition time node. For example, if the acquisition time node of the current user information is December 2023, then the above-mentioned user annual information is all user information from December 2022 to December 2023.

[0090] In this embodiment, the regression prediction model to be trained can be an existing machine learning model such as XGBoost (Extreme Gradient Boosting), Random Forest or ARIMA (Autoregressive Integrated Moving Average Model) downloaded from the Internet.

[0091] In this embodiment, the above-mentioned annual time characteristics may include but are not limited to purchase frequency, user's purchase behavior in a specific season or month, transaction time interval (i.e., the average time between two consecutive purchases by the user), the first and last purchase dates of the user within a year, and the average amount of user transactions within a certain time period.

[0092] In a possible embodiment, the time node for information acquisition can be determined, and user information from the year before the time node can be extracted from the historical transaction database as a user training set. The above user annual information is subjected to time series analysis to extract key time node features, such as transaction frequency, seasonal purchase pattern, average transaction amount, etc. The above annual time features are used to estimate model training parameters. The above model training parameters include but are not limited to regularization coefficient, maximum tree depth, learning rate, etc. The above model training parameters are configured to a gradient boosting machine to be trained (i.e., the above regression prediction model to be trained), and the model is trained using the obtained user training set. Through iterative optimization, the model parameters are adjusted to minimize the prediction error, and finally a fine-tuned regression prediction model that can accurately predict user retention rate is obtained.

[0093] This application comprehensively analyzes the user's annual transaction data and behavioral characteristics, extracts key time node features, and uses these features to train a regression prediction model, so that the regression prediction model can more accurately predict the user's customer lifetime value and identify potential high-value or high-risk customers.

[0094] In some optional implementations, after the above step of “importing the model training parameters into the regression prediction model to be trained, and training the regression prediction model to be trained using the user training set to obtain a preset regression prediction model”, the following steps are included:

[0095] Obtain a user test set from the historical transaction database, import the user test set into a preset regression prediction model, and output the prediction result;

[0096] Based on the prediction results and the preset standard results, the back propagation algorithm is used for fitting to obtain the prediction error;

[0097] The prediction error is compared with a preset threshold. If the prediction error is greater than the preset threshold, the preset regression prediction model is iteratively updated until the prediction error is less than or equal to the preset threshold, and the preset regression prediction model is updated.

[0098] In this embodiment, a user test set can be extracted from the historical transaction database and input into the trained regression prediction model to generate preliminary prediction results. Next, the model is fitted using a back propagation algorithm. The back propagation algorithm can minimize the prediction error by calculating the error between the predicted value and the actual value and propagating it back to the model to adjust the weight. The obtained prediction error is compared with a preset threshold: if the error exceeds the threshold, the model parameters of the regression prediction model are iteratively updated, such as adjusting the learning rate, modifying the depth of the decision tree, or adjusting other related parameters, and then the prediction and fitting process is repeated until the prediction error is lower than or equal to the threshold.

[0099] This application can systematically reduce errors and improve prediction accuracy by inputting user test set data into a trained model and outputting prediction results, and then using a back-propagation algorithm to correct the prediction errors. This improves the prediction performance of the model, enhances the model's adaptability and generalization ability when facing new data, and makes it a more reliable and effective prediction tool.

[0100] In some optional implementations, step S203 includes the following steps:

[0101] The policy information is extracted through the input layer to obtain the policy feature tensor;

[0102] The processing layer calculates the lifetime value of the policy feature tensor and the user retention rate to obtain the customer lifetime value;

[0103] Output customer lifetime value through the output layer.

[0104] In this embodiment, the input layer is the first layer of the regression prediction model, which is responsible for receiving raw data input. Specifically, the input layer processes policy information, such as policy type, amount, term, customer's historical transaction data, etc. The input layer can also perform feature extraction processing to convert the raw data into a format that the model can process, such as one-hot encoding categorical data or standardizing continuous variables.

[0105] In this embodiment, the processing layer is located between the input layer and the output layer in the neural network. There can be one or more processing layers, which are responsible for performing complex data conversion and feature extraction. Specifically, the processing layer will combine the policy feature tensor and the user retention rate to calculate the customer's lifetime value (CLV). The processing layer can perform operations such as weighting and activation functions.

[0106] In this embodiment, the output layer is the last layer of the neural network, which is responsible for outputting the prediction results of the model. Specifically, the output layer will display the calculated customer lifetime value. The design of the output layer may depend on the requirements of the specific task. For example, a regression problem may use a single node to output continuous values, while a classification problem may use multiple nodes to output class probabilities.

[0107] In this embodiment, the above life cycle value calculation can be implemented by the following formula:

[0108]

[0109] Among them, the above P is the total number of insurance policies held by the customer, the above c is the cost of the pth policy (including costs such as claims and operating expenses), the above t is the time node, the above F is the premium, and the above r is the customer retention rate at time t.

[0110] In a possible embodiment, after obtaining the customer retention probability data, the time data of the past year can be used as feature X, and the customer lifetime value at each time node can be used as Y. The above feature X and feature Y can be input into the regression model as correction parameters for secondary training, so as to improve the accuracy of predicting the customer lifetime value at time node t.

[0111] This application can improve the efficiency of data processing through life cycle value calculation, and also enhance the accuracy and reliability of the prediction model, providing a stable foundation for subsequent risk management and marketing strategy formulation.

[0112] In some optional implementations, step S204 includes the following steps:

[0113] Randomly sample the customer lifetime value to obtain the user value path;

[0114] The user value path is predicted and distributed through the Bayesian inference method to obtain the user risk assessment.

[0115] In this embodiment, the above-mentioned Bayesian inference method can be a statistical reasoning method based on Bayes' theorem. Specifically, assuming that an insurance company wants to predict the long-term value of a new customer, it can establish a prior probability distribution by analyzing historical customer data (such as age, gender, policy type, claim history, etc.). When a new customer purchases insurance, the company can use the initial transaction data of these customers to update the prediction of the customer's lifetime value. Through Bayesian inference, the company can combine prior knowledge and newly acquired data to calculate the posterior distribution, thereby more accurately predicting the long-term value and risk level of new customers.

[0116] In this embodiment, the user value path may be a sequence of changes in the value of a customer throughout the entire period of interaction with the company. The user value path may include various indicators of customer-generated revenue, transaction frequency, purchase pattern, service utilization, and customer loyalty, and the user value path may reflect the value estimate from the first time of becoming a customer to the current or predicted future time points.

[0117] In one possible embodiment, the accumulated customer data is first randomly sampled to extract user value paths representing different customer behaviors and value evolutions. The user value paths include various economic indicators, such as consumption frequency, purchase volume, service utilization, and customer loyalty. Next, the Bayesian inference method is used to analyze these paths, and a probability model is constructed to estimate customer behaviors and values ​​in various future scenarios to obtain user risk assessment.

[0118] By using the Bayesian inference method to perform random sampling and risk assessment on customer lifetime value, this application enables insurance companies to achieve an in-depth understanding and accurate prediction of customer behavior and value changes, allowing insurance companies to identify and quantify risks and opportunities, thereby optimizing customer relationship management and resource allocation.

[0119] In some optional implementations, after the above step of "performing a predictive distribution process on the user value path by using the Bayesian inference method to obtain a user risk assessment", the following steps are also included:

[0120] Divide the user groups by value based on user risk assessment to obtain the target user value groups;

[0121] Carry out differentiated strategic delivery to user value groups to obtain delivery strategies.

[0122] In this embodiment, the target user value group may include high-value customers and high-risk customers. Specifically, the high-value customers may be customers who make the greatest long-term contribution to the enterprise. The criteria for judging the high-value customers may include, but are not limited to, evaluation indicators such as high purchase frequency, large purchase volume, long-term loyalty, and high satisfaction with the enterprise's products and services; the high-risk users may be customers who have a high probability of reducing purchase frequency, reducing consumption, or completely losing in the near future. The criteria for judging the high-risk customers may include, but are not limited to, evaluation indicators such as transaction reduction trend, service complaint record, and increased price sensitivity.

[0123] In this embodiment, the above-mentioned investment strategy can be to increase investment in high-value customers to maximize the company's profits, and reduce investment in high-risk customers. If high-risk customers are lost, the company may reduce losses and achieve the purpose of optimizing the customer structure.

[0124] In one possible embodiment, the customer's life cycle value and behavioral risk can be deeply evaluated through integrated analysis tools such as Apache Kafka and Apache Flink, and data mining and statistical analysis techniques such as association rule learning, classification, and clustering can be used to divide customers into different value groups, such as high-value customers, medium-value customers, and high-risk customers. Then, based on the specific characteristics and needs of the above value groups, differentiated marketing strategies can be designed and implemented. For example, for high-value customer groups, value-added services, loyalty rewards, and personalized communication strategies can be implemented to enhance customer satisfaction and retention; for high-risk customers, the churn rate can be reduced through customized offers, special attention, and services that meet their specific needs. This process involves cross-departmental collaboration, including marketing, customer service, and product development teams, to ensure the effective implementation of the strategy and monitor its effectiveness, thereby maximizing ROI and improving overall customer satisfaction and corporate competitiveness.

[0125] This application divides user groups into different groups based on user risk assessment and implements differentiated strategy delivery to identify high-value customers and high-risk customers, thereby designing personalized services and marketing strategies for different customer groups, enabling insurance companies to achieve refined customer management and optimized resource allocation, thereby improving customer satisfaction and loyalty, optimizing the investment efficiency of marketing budgets, and enhancing the market competitiveness and overall profitability of insurance companies.

[0126] It is understandable that in the specific implementation of this application, data related to transaction history data sets, user retention rates, customer lifetime value, and user risk assessment are involved. When the embodiments in this application are applied to specific products or technologies, user permission or consent is required, and the collection, use, and processing of relevant data must comply with relevant laws, regulations, and standards of relevant countries and regions.

[0127] It should be emphasized that in order to further ensure the privacy and security of the above-mentioned transaction history data set, user retention rate, customer lifetime value, user risk assessment and other information, the above-mentioned transaction history data set, user retention rate, customer lifetime value, user risk assessment and other information can also be stored in a blockchain node.

[0128] The blockchain referred to in this application is a new application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm, etc. Blockchain is essentially a decentralized database, a string of data blocks generated by cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify the validity of its information (anti-counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, platform product service layer, and application service layer.

[0129] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0130] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0131] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0132] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0133] Further references Figure 3 , as a response to the above Figure 2 In order to realize the method shown in the figure, the present application provides an embodiment of a risk assessment device, and the device embodiment is Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0134] like Figure 3As shown, the risk assessment device 300 described in this embodiment includes: an acquisition module 301, a processing module 302, a calculation module 303 and an analysis module 304. Among them:

[0135] An acquisition module 301 is used to acquire a transaction history data set, wherein the transaction history data set consists of policy information and user information;

[0136] The processing module 302 is used to perform retention prediction processing on the user information to obtain a user retention rate;

[0137] A calculation module 303 is used to input the user retention rate and the policy information into a preset regression prediction model for calculation to obtain a customer life cycle value;

[0138] The analysis module 304 is used to perform simulated sampling analysis on the customer life cycle value to obtain a user risk assessment.

[0139] In a possible embodiment, the processing module 302 specifically includes:

[0140] The first processing submodule is used to process the user information to obtain a user transaction factor;

[0141] The second processing submodule inputs the user transaction factor into a preset LGBM model for prediction processing to obtain the user retention rate.

[0142] In a possible embodiment, before the calculation module 303, the method further includes:

[0143] A time acquisition module, used to acquire the acquisition time node of the user information and the regression prediction model to be trained;

[0144] A pulling module, used to pull the user annual information of the year before the acquisition time node from the historical transaction database as a user training set;

[0145] An extraction module, used to extract time nodes from the user's annual information to obtain annual time features;

[0146] An estimation module, used to estimate the annual time characteristics and obtain model training parameters;

[0147] The training module is used to import the model training parameters into the regression prediction model to be trained, and use the user training set to train the regression prediction model to be trained to obtain a preset regression prediction model.

[0148] In a possible embodiment, after the training module, the following is further included:

[0149] An import module, used to obtain a user test set from a historical transaction database, import the user test set into the preset regression prediction model, and output a prediction result;

[0150] The fitting module is used to fit the prediction results and the preset standard results using the back propagation algorithm to obtain the prediction error;

[0151] The comparison module is used to compare the prediction error with a preset threshold. If the prediction error is greater than the preset threshold, the preset regression prediction model is iteratively updated until the prediction error is less than or equal to the preset threshold, and the preset regression prediction model is updated.

[0152] In a possible embodiment, the calculation module 303 specifically includes:

[0153] An extraction submodule, used for performing feature extraction on the policy information through the input layer to obtain a policy feature tensor;

[0154] A calculation submodule, configured to calculate the lifetime value of the policy feature tensor and the user retention rate through the processing layer to obtain the customer lifetime value;

[0155] An output submodule is used to output the customer lifetime value through the output layer.

[0156] In a possible embodiment, the analysis module 304 specifically includes:

[0157] A sampling module is used to randomly sample the customer life cycle value to obtain a user value path;

[0158] The prediction distribution submodule is used to perform prediction distribution processing on the user value path through Bayesian inference method to obtain the user risk assessment.

[0159] In a possible embodiment, after the prediction distribution submodule, the following is further included:

[0160] A division submodule is used to divide the user groups into value groups based on the user risk assessment to obtain target user value groups;

[0161] The delivery submodule is used to carry out differentiated delivery of strategies to the user value groups to obtain delivery strategies.

[0162] In this embodiment, by obtaining a transaction history data set, the transaction history data set consists of policy information and user information, and performing retention prediction processing on the user information to obtain a user retention rate. The user retention rate and the policy information are input into a preset regression prediction model for calculation to obtain a customer lifetime value. A simulated sampling analysis is performed on the customer lifetime value to obtain a user risk assessment. The customer lifetime value is calculated through retention prediction and regression prediction models, and risk assessment is performed through simulated sampling analysis, so that insurance companies can more accurately identify and distinguish high-risk and low-risk customers, thereby achieving optimal allocation of resources and precise management of risks, so that insurance companies can optimize customer structure and formulate more targeted customer maintenance and market strategies, thereby maximizing the benefits of insurance companies with limited resources.

[0163] In this embodiment, the operations performed by the above units or modules correspond one-to-one to the steps of the risk assessment method in the above implementation mode, which will not be described in detail here.

[0164] To solve the above technical problems, the present application also provides a computer device. Figure 4 , Figure 4 This is a basic structural block diagram of the computer device in this embodiment.

[0165] The computer device 4 includes a memory 41, a processor 42, and a network interface 43 that are interconnected through a system bus. It should be noted that the figure only shows a computer device 4 with components 41-43, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations 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 (Application Specific Integrated Circuit, ASIC), programmable gate arrays (Field-Programmable Gate Array, FPGA), digital processors (Digital Signal Processor, DSP), embedded devices, etc.

[0166] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device.

[0167] The memory 41 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, 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 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 equipped on the computer device 4, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), etc. Of course, the memory 41 can also include both the internal storage unit of the computer device 4 and its external storage device. In this embodiment, the memory 41 is generally used to store the operating system and various application software installed on the computer device 4, such as computer-readable instructions of the risk assessment method, etc. In addition, the memory 41 can also be used to temporarily store various types of data that have been output or are to be output.

[0168] The processor 42 may 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 for running the risk assessment method.

[0169] The network interface 43 may include a wireless network interface or a wired network interface. The network interface 43 is generally used to establish a communication connection between the computer device 4 and other electronic devices.

[0170] The present embodiment provides a computer device, which obtains a transaction history data set, wherein the transaction history data set consists of policy information and user information, performs retention prediction processing on the user information to obtain a user retention rate, inputs the user retention rate and the policy information into a preset regression prediction model for calculation to obtain a customer lifetime value, performs simulated sampling analysis on the customer lifetime value to obtain a user risk assessment, calculates the customer lifetime value through retention prediction and regression prediction models, and performs risk assessment through simulated sampling analysis, so that insurance companies can more accurately identify and distinguish high-risk and low-risk customers, thereby achieving optimal allocation of resources and precise management of risks, so that insurance companies can optimize customer structure and formulate more targeted customer maintenance and market strategies, thereby maximizing the benefits of insurance companies with limited resources.

[0171] The present application also provides another embodiment, namely, providing a computer-readable storage medium, wherein the computer-readable storage medium stores computer-readable instructions, and the computer-readable instructions can be executed by at least one processor to enable the at least one processor to perform the steps of the risk assessment method as described above.

[0172] The present embodiment provides a computer-readable storage medium, which obtains a transaction history data set, wherein the transaction history data set consists of policy information and user information, performs retention prediction processing on the user information to obtain a user retention rate, inputs the user retention rate and the policy information into a preset regression prediction model for calculation to obtain a customer lifetime value, performs simulated sampling analysis on the customer lifetime value to obtain a user risk assessment, calculates the customer lifetime value through retention prediction and regression prediction models, and performs risk assessment through simulated sampling analysis, so that insurance companies can more accurately identify and distinguish high-risk and low-risk customers, thereby achieving optimal allocation of resources and precise management of risks, allowing insurance companies to optimize customer structure and formulate more targeted customer maintenance and market strategies, thereby maximizing the benefits of insurance companies with limited resources.

[0173] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0174] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.

Claims

1. A risk assessment method, characterized in that: The steps include: Acquire a transaction history data set, wherein the transaction history data set consists of policy information and user information; Performing retention prediction processing on the user information to obtain a user retention rate; Input the user retention rate and the policy information into a preset regression prediction model for calculation to obtain the customer life cycle value; A simulated sampling analysis is performed on the customer lifetime value to obtain a user risk assessment.

2. The risk assessment method according to claim 1, characterized in that: The step of performing retention prediction processing on the user information to obtain the user retention rate specifically includes: Performing factor processing on the user information to obtain a user transaction factor; The user transaction factor is input into a preset LGBM model for prediction processing to obtain the user retention rate.

3. The risk assessment method according to claim 1, characterized in that: Before the step of inputting the user retention rate and the policy information into a preset regression prediction model for calculation to obtain the customer life cycle value, the method further includes: Obtaining the acquisition time node of the user information and the regression prediction model to be trained; Pull out the user annual information of the year before the acquisition time node from the historical transaction database as the user training set; Extracting time nodes from the user's annual information to obtain annual time features; Estimating the annual time characteristics to obtain model training parameters; The model training parameters are imported into the regression prediction model to be trained, and the regression prediction model to be trained is trained using the user training set to obtain a preset regression prediction model.

4. The risk assessment method according to claim 3, characterized in that: After the step of importing the model training parameters into the regression prediction model to be trained, and training the regression prediction model to be trained using the user training set to obtain a preset regression prediction model, the method further includes: Obtaining a user test set from a historical transaction database, and importing the user test set into the preset regression prediction model, and outputting a prediction result; Based on the prediction results and the preset standard results, the back propagation algorithm is used for fitting to obtain the prediction error; The prediction error is compared with a preset threshold. If the prediction error is greater than the preset threshold, the preset regression prediction model is iteratively updated until the prediction error is less than or equal to the preset threshold, and the preset regression prediction model is updated.

5. The risk assessment method according to claim 1, characterized in that: The preset regression prediction model includes an input layer, a processing layer and an output layer. The step of inputting the user retention rate and the policy information into the preset regression prediction model for calculation to obtain the customer life cycle value specifically includes: Extracting features of the policy information through the input layer to obtain a policy feature tensor; The processing layer calculates the lifetime value of the policy feature tensor and the user retention rate to obtain the customer lifetime value; The customer lifetime value is outputted through the output layer.

6. The risk assessment method according to claim 1, characterized in that: The step of performing simulated sampling analysis on the customer life cycle value to obtain a user risk assessment specifically includes: Randomly sampling the customer life cycle value to obtain a user value path; The user value path is processed for predictive distribution through Bayesian inference method to obtain the user risk assessment.

7. The risk assessment method according to claim 1, characterized in that: After the step of performing a simulated sampling analysis on the customer life cycle value to obtain a user risk assessment, the method further includes: Based on the user risk assessment, the user groups are divided into value groups to obtain target user value groups; A differentiated strategy is deployed for the user value groups to obtain a deployment strategy.

8. A risk assessment device, characterized in that: include: An acquisition module is used to acquire a transaction history data set, wherein the transaction history data set consists of policy information and user information; A processing module, used to perform retention prediction processing on the user information to obtain a user retention rate; A calculation module, used for inputting the user retention rate and the policy information into a preset regression prediction model for calculation to obtain the customer life cycle value; The analysis module is used to perform simulated sampling analysis on the customer life cycle value to obtain a user risk assessment.

9. A computer device, characterized in that: The method comprises a memory and a processor, wherein the memory stores computer-readable instructions, and the processor implements the steps of the risk assessment method according to any one of claims 1 to 7 when executing the computer-readable instructions.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by a processor, the steps of the risk assessment method according to any one of claims 1 to 7 are implemented.

Images