Insurance user value estimation method and device, electronic equipment and medium
By training the user value estimation model for holidays and non-holidays respectively, and calculating the user value coefficient to correct the initial estimation data, the problems of long model training period and inaccurate estimation in the existing technology are solved, and the problem of insured users is realized. Accurate estimation of value.
Patent Information
- Application Number
- CN202411381443.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-30
- Publication Date
- 2025-01-21
AI Technical Summary
In the prior art, holiday data is included as features into the model, making the model training cycle long and it is impossible to accurately estimate the value of insured users.
By obtaining the target insurance policy data of the target user, the insurance time is determined, and user value estimation models for holidays and non-holidays are trained based on the insurance time. Input target insurance policy data into these models, calculate user value coefficients, and correct the initial estimated data to obtain more accurate user value estimates.
It achieves accurate estimation and identification of user value during different types of holidays, shortens the model training cycle, improves the accuracy of insured user value estimation, and helps business data decision-making.
Smart Images
Figure CN119338598A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of data processing and model estimation, and in particular to an insurance user value estimation method, device, electronic equipment and medium. Background Art
[0002] In recent years, with the rise of Internet insurance brokerage platforms, short-term insurance products, mainly medical and accidental insurance, have occupied an increasing share. Compared with traditional insurance sales channels, Internet insurance brokerage platforms can obtain more user-side behavioral data, thereby using big data, machine learning and other technologies to estimate the value of insured users.
[0003] For short-term insurance products, users’ performance on holidays is different from that on non-holidays, and their performance may be different during different holidays (such as May Day, National Day, Spring Festival, other holidays, etc.). Traditional estimation models generally incorporate holiday data as features into the model, which will prolong the model training cycle and fail to accurately capture recent user changes. In addition, the performance of historical holiday users may be different from that of recent holiday users, and it is impossible to fully reflect the recent user holiday performance. Therefore, it is impossible to accurately estimate the value of insured users. Summary of the invention
[0004] The present invention provides an insurance user value estimation method, device, electronic device and medium, which are used to solve the defects in the prior art that incorporating holiday data as features into the model makes the model training cycle long and cannot accurately estimate the insured user value, and realizes accurate estimation and identification of different types of holiday user values, and assists business data decision-making.
[0005] The present invention provides an insurance user value estimation method, comprising: Acquire target insurance policy data of a target user and determine the insurance time based on the target insurance policy data; Determine a target insurance user value estimation model based on the insurance time, the target insurance user value estimation model includes a target holiday insurance user value estimation model and a target non-holiday insurance user value estimation model, the target holiday insurance user value estimation model and the target non-holiday insurance user value estimation model are obtained by training the user insurance policy data during the target holiday period and the user insurance policy data before and after the target holiday; Inputting the target insurance policy data into the initial insurance user value estimation model, the target holiday insurance user value estimation model and the target non-holiday insurance user value estimation model respectively, and obtaining the first estimation data, the second estimation data and the third estimation data outputted by the models respectively; Calculate a user value coefficient based on the second estimated data and the third estimated data; The first estimated data is corrected by using the user value coefficient to obtain a target estimated insurance user value.
[0006] In a possible implementation, the method further includes: The user value coefficient is calculated by a first formula based on the second estimated data and the third estimated data, wherein the first formula is: y=x1 / x2; Among them, y is the user value coefficient, x1 is the second estimated data, and x2 is the third estimated data.
[0007] In a possible implementation, the method further includes: Based on a second formula, the first estimated data is corrected by the user value coefficient to obtain a target estimated insurance user value, wherein the second formula is: L=z×y; Among them, L is the target estimated insurance user value, z is the first estimated data, and y is the user value coefficient.
[0008] In a possible implementation, the method further includes: Determine the corresponding target holiday based on the insurance time; A target holiday insurance user value estimation model and a target non-holiday insurance user value estimation model are determined based on the target holidays.
[0009] In a possible implementation, the method further includes: The target insurance user value estimation model is trained by the following steps: Obtaining the insurance policy data of the first user during the target holiday and the insurance policy data of the second user before and after the target holiday; Performing data feature processing on the first user insurance policy data and the second user insurance policy data to obtain first sample user insurance policy data and second sample user insurance policy data after data processing; Dividing the first sample user insurance policy data and the second sample user insurance policy data into a model training set and a model test set respectively; Training a holiday insurance user value estimation model using the model training set in the first sample user's insurance policy data, and training a non-holiday insurance user value estimation model using the model training set in the second sample user's insurance policy data; The holiday insurance user value estimation model after initial training is tested by the model test set in the insurance policy data of the first sample user to obtain a first test result, and the non-holiday insurance user value estimation model after initial training is tested by the model test set in the insurance policy data of the second sample user to obtain a second test result; When the first test result and the second test result meet the preset conditions, it is determined that the training of the holiday insurance user value estimation model and the non-holiday insurance user value estimation model is completed, and the target insurance user value estimation model is obtained.
[0010] In a possible implementation, the method further includes: The test results shall at least include the model AUC evaluation value, estimated accuracy and model recall rate; Determine whether the AUC evaluation value, the estimated accuracy rate, and the model recall rate in the first test result and the second test result respectively reach corresponding evaluation thresholds; If so, it is determined that the training of the holiday insurance user value estimation model and the non-holiday insurance user value estimation model is completed, and the target insurance user value estimation model is obtained; If not, the model parameters are modified and the holiday insurance user value estimation model is re-trained using the model training set in the first sample user's insurance policy data, and the non-holiday insurance user value estimation model is trained using the model training set in the second sample user's insurance policy data.
[0011] The present invention also provides an insurance user value estimation device, comprising the following modules: An acquisition module, used to acquire target insurance policy data of a target user and determine the insurance time based on the target insurance policy data; A determination module, configured to determine a target insurance user value estimation model based on the insurance time, wherein the target insurance user value estimation model includes a target holiday insurance user value estimation model and a target non-holiday insurance user value estimation model; An estimation module, used for inputting the target insurance policy data into the initial insurance user value estimation model, the target holiday insurance user value estimation model and the target non-holiday insurance user value estimation model, respectively, to obtain first estimation data, second estimation data and third estimation data outputted by the models respectively; The estimation module is further used to calculate a user value coefficient based on the second estimation data and the third estimation data; The estimation module is further used to correct the first estimation data by using the user value coefficient to obtain a target estimated insurance user value.
[0012] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the insurance user value estimation method as described in any one of the above is implemented.
[0013] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described insurance user value estimation methods.
[0014] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the insurance user value estimation method as described in any one of the above is implemented.
[0015] The insurance user value estimation method, device, electronic device and medium provided by the present invention obtain the target insurance policy data of the target user and determine the insurance time based on the target insurance policy data; determine the target insurance user value estimation model based on the insurance time, and the target insurance user value estimation model includes a target holiday insurance user value estimation model and a target non-holiday insurance user value estimation model; input the target insurance policy data into the initial insurance user value estimation model, the target holiday insurance user value estimation model and the target non-holiday insurance user value estimation model respectively, and obtain the first estimation data, the second estimation data and the third estimation data output by the models respectively; calculate the user value coefficient based on the second estimation data and the third estimation data; correct the first estimation data by the user value coefficient to obtain the target estimated insurance user value. Compared with the existing technology that incorporates holiday data as features into the model, which makes the model training cycle long and cannot accurately estimate the value of insured users, this solution trains models according to different holiday types, thereby calculating the user value coefficients of different types of holidays. On the basis of the original user value estimation model, the user value data is corrected according to the corresponding holiday type, making the user value estimation more accurate and facilitating business data decision-making. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0017] Figure 1 It is a flow chart of the insurance user value estimation method provided by the present invention.
[0018] Figure 2It is a flow chart of the target insurance user value estimation model training method provided by the present invention.
[0019] Figure 3 It is a schematic flow chart of the working principle of the insurance user value estimation method provided by the present invention.
[0020] Figure 4 It is a structural schematic diagram of the insurance user value estimation device provided by the present invention.
[0021] Figure 5 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION
[0022] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0023] To facilitate understanding of the embodiments of the present invention, specific embodiments will be further explained below in conjunction with the accompanying drawings. The embodiments do not constitute a limitation on the embodiments of the present invention.
[0024] Figure 1 is a flow chart of the insurance user value estimation method provided by the present invention, such as Figure 1 As shown, the method specifically includes: S11. Obtain target insurance policy data of a target user and determine an insurance application time based on the target insurance policy data.
[0025] In the embodiment of the present invention, combined with Figure 3 The working principle flow chart of the insurance user value estimation method shown in the figure is described in detail: First, the target insurance policy data of the target user is obtained. The target insurance policy data also includes the corresponding user data, including but not limited to basic user characteristics (age, gender, province, city, etc.), user behavior data (such as browsing pages, purchasing policies, refunding policies, policy purchase time and page browsing time, etc.). The insurance time can be determined based on the time of purchasing the policy.
[0026] S12: Determine a target insurance user value estimation model based on the insurance time.
[0027] In the embodiment of the present invention, first, holidays can be divided into four types: May Day, National Day, Spring Festival and other holidays. For each holiday, a corresponding holiday insurance user value estimation model and a non-holiday insurance user value estimation model are pre-trained. The model training method is Figure 2 The details are described in the corresponding embodiments.
[0028] Furthermore, the corresponding target holidays can be determined by querying the calendar according to the insurance time; based on the target holidays, the target insurance user value estimation model can be determined, and the target insurance user value estimation model includes a target holiday insurance user value estimation model and a target non-holiday insurance user value estimation model. The target holiday insurance user value estimation model and the target non-holiday insurance user value estimation model are trained through user insurance policy data during the target holidays and user insurance policy data before and after the target holidays.
[0029] For example, if the insurance purchase date is May 2, the corresponding target holiday can be determined to be May Day. Then, the target insurance user value estimation model can be determined, including the target holiday insurance user value estimation model corresponding to May Day and the target non-holiday insurance user value estimation model.
[0030] S13, inputting the target insurance policy data into the initial insurance user value estimation model, the target holiday insurance user value estimation model and the target non-holiday insurance user value estimation model respectively, and obtaining the first estimation data, the second estimation data and the third estimation data outputted by the models respectively.
[0031] S14. Calculate a user value coefficient based on the second estimated data and the third estimated data.
[0032] The target insurance policy data is respectively input into the initial insurance user value estimation model, the target holiday insurance user value estimation model and the target non-holiday insurance user value estimation model to obtain the first estimation data, the second estimation data and the third estimation data outputted by the model respectively. The initial insurance user value estimation model is an existing model and is not specifically limited in the embodiment of the present invention.
[0033] Furthermore, the user value coefficient is calculated based on the second estimated data and the third estimated data by the first formula, and the first formula is: y=x1 / x2; wherein y is the user value coefficient, x1 is the second estimated data, and x2 is the third estimated data.
[0034] S15. Correct the first estimated data using the user value coefficient to obtain a target estimated insurance user value.
[0035] Based on the second formula, the first estimated data is corrected by the user value coefficient to obtain the target estimated insurance user value. The second formula is: L=z×y; wherein L is the target estimated insurance user value, z is the first estimated data, and y is the user value coefficient.
[0036] Furthermore, the user value estimation results can be used to formulate targeted marketing strategies and customer relationship management strategies. During holidays, personalized services can be provided to high-value users.
[0037] The insurance user value estimation method provided by the present invention obtains the target insurance policy data of the target user and determines the insurance time based on the target insurance policy data; determines the target insurance user value estimation model based on the insurance time, and the target insurance user value estimation model includes a target holiday insurance user value estimation model and a target non-holiday insurance user value estimation model; inputs the target insurance policy data into the initial insurance user value estimation model, the target holiday insurance user value estimation model and the target non-holiday insurance user value estimation model, respectively, to obtain first estimation data, second estimation data and third estimation data output by the models respectively; calculates the user value coefficient based on the second estimation data and the third estimation data; corrects the first estimation data by the user value coefficient to obtain the target estimated insurance user value. Compared with the existing technology that incorporates holiday data as features into the model, which makes the model training cycle long and cannot accurately estimate the value of insured users, this method trains models according to different holiday types, thereby calculating the user value coefficients of different types of holidays. On the basis of the original user value estimation model, the user value data is corrected according to the corresponding holiday type, making the user value estimation more accurate and facilitating business data decision-making.
[0038] Figure 2 is a flow chart of the target insurance user value estimation model training method provided by the present invention, such as Figure 2 As shown, the method specifically includes: S21. Obtain insurance policy data of a first user during a target holiday and insurance policy data of a second user before and after the target holiday.
[0039] In the embodiment of the present invention, firstly, holidays can be divided into four types: May Day, National Day, Spring Festival and other holidays. This classification helps to conduct a more detailed analysis of the characteristics of different holidays.
[0040] For each holiday, the first user's insurance policy data during the holiday and the second user's insurance policy data before and after the holiday are collected. For each insurance policy data, the corresponding user data can also be collected, including but not limited to user basic characteristics (age, gender, occupation, province, city, etc.), user behavior data (such as browsing pages, purchasing insurance policies, refunding policies, purchasing policy time points and browsing page time, etc.).
[0041] S22: Perform data feature processing on the first user insurance policy data and the second user insurance policy data to obtain first sample user insurance policy data and second sample user insurance policy data after data processing.
[0042] Perform data feature processing on the first user's insurance policy data and the second user's insurance policy data. Analyze changes in user behavior before and after holidays to identify key factors that affect user value. For example, perform feature engineering on the data, including feature selection, feature extraction, and feature conversion. The user's purchase frequency and purchase amount can be converted into indicators of user value.
[0043] Data cleaning is the processing of data before training the model, such as: missing data filling, data outlier processing, format and data type conversion, data normalization, etc. The data can also be normalized to eliminate the influence of different dimensions. The first sample user insurance policy data and the second sample user insurance policy data are obtained after data processing.
[0044] S23. Divide the first sample user insurance policy data and the second sample user insurance policy data into a model training set and a model testing set respectively.
[0045] In the embodiment of the present invention, the division ratio can be set, for example, 80% of the sample data is used as the model training set, and the remaining 20% is used as the model test set. The data can be randomly selected during the division.
[0046] S24. Train a holiday insurance user value estimation model using the model training set in the first sample user's insurance policy data, and train a non-holiday insurance user value estimation model using the model training set in the second sample user's insurance policy data.
[0047] The model training set of the first sample user's insurance policy data and the model training set of the second sample user's insurance policy data are labeled, the user value corresponding to each policy is labeled, and then the model training set is input into the initial model for training. Machine learning algorithms such as random forest, gradient boosting machine (GBM) or deep learning models can be used.
[0048] Furthermore, the trained model is cross-validated to evaluate the generalization ability of the model. The model parameters are adjusted and feature importance analysis is performed to optimize the model performance.
[0049] S25. The holiday insurance user value estimation model after initial training is tested by the model test set in the insurance policy data of the first sample user to obtain a first test result, and the non-holiday insurance user value estimation model after initial training is tested by the model test set in the insurance policy data of the second sample user to obtain a second test result.
[0050] The test results of the embodiment of the present invention at least include the model AUC evaluation value, the estimation accuracy and the model recall rate. The model test set in the first sample user insurance policy data is input into the holiday insurance user value estimation model after initial training to obtain the first test result, and the model test set in the second sample user insurance policy data is input into the non-holiday insurance user value estimation model after initial training to obtain the second test result.
[0051] S26. When the first test result and the second test result meet the preset conditions, it is determined that the training of the holiday insurance user value estimation model and the non-holiday insurance user value estimation model is completed, and the target insurance user value estimation model is obtained.
[0052] Determine whether the AUC evaluation value, estimation accuracy and model recall rate in the first test result and the second test result reach the corresponding evaluation thresholds respectively; if so, determine that the training of the holiday insurance user value estimation model and the non-holiday insurance user value estimation model is completed, and obtain the target insurance user value estimation model; if not, modify the model parameters and re-train the holiday insurance user value estimation model through the model training set in the first sample user insurance policy data, and train the non-holiday insurance user value estimation model through the model training set in the second sample user insurance policy data.
[0053] For example, the embodiments of the present invention use a tree model, and generally first measure the model's AUC (Area Under Curve) value (>0.6) as well as accuracy, recall rate, etc., and compare the difference between the estimated data output by the training model and the data label. The smaller the difference, the better the model training effect.
[0054] Optionally, you can also use actual user value data to evaluate the model, such as using indicators such as mean square error (MSE) and absolute error (MAE). Adjust and optimize the model based on the evaluation results. Update the model regularly to adapt to changes in user behavior and new holiday data.
[0055] Furthermore, the target model can be integrated into the existing user management system to achieve automated user value estimation. A user interface is provided so that business analysts and decision makers can easily access the estimation results and model performance. The user value estimation results can be used to formulate targeted marketing strategies and customer relationship management strategies. During holidays, personalized services can be provided to high-value users.
[0056] The target insurance user value estimation model training method provided by the present invention comprises the following steps: obtaining the insurance policy data of the first user during the target holiday and the insurance policy data of the second user before and after the target holiday; performing data feature processing on the first user insurance policy data and the second user insurance policy data to obtain the first sample user insurance policy data and the second sample user insurance policy data after data processing; dividing the first sample user insurance policy data and the second sample user insurance policy data into a model training set and a model test set respectively; training the holiday insurance user value estimation model through the model training set in the first sample user insurance policy data, and training the non-holiday insurance user value estimation model through the model training set in the second sample user insurance policy data; testing the holiday insurance user value estimation model after initial training through the model test set in the first sample user insurance policy data to obtain a first test result, and testing the non-holiday insurance user value estimation model after initial training through the model test set in the second sample user insurance policy data to obtain a second test result; when the first test result and the second test result meet the preset conditions, it is determined that the training of the holiday insurance user value estimation model and the non-holiday insurance user value estimation model is completed, and the target insurance user value estimation model is obtained. Compared with the existing technology that incorporates holiday data as features into the model, which makes the model training cycle long and cannot accurately estimate the value of insured users, this method trains models according to different holiday types, thereby calculating the user value coefficients of different types of holidays, and corrects the user value data according to the corresponding holiday types on the basis of the original user value estimation model, making the user value estimation more accurate.
[0057] The insurance user value estimation device provided by the present invention is described below. The insurance user value estimation device described below and the insurance user value estimation method described above can be referenced to each other.
[0058] Figure 4 : is a schematic diagram of the structure of the insurance user value estimation device provided by the present invention, which specifically includes: An acquisition module 401 is used to acquire target insurance policy data of a target user and determine the insurance time based on the target insurance policy data; A determination module 402 is used to determine a target insurance user value estimation model based on the insurance time, wherein the target insurance user value estimation model includes a target holiday insurance user value estimation model and a target non-holiday insurance user value estimation model, wherein the target holiday insurance user value estimation model and the target non-holiday insurance user value estimation model are obtained by training the user insurance policy data during the target holiday period and the user insurance policy data before and after the target holiday period; The estimation module 403 is used to input the target insurance policy data into the initial insurance user value estimation model, the target holiday insurance user value estimation model and the target non-holiday insurance user value estimation model, respectively, to obtain the first estimation data, the second estimation data and the third estimation data outputted by the models respectively; The estimation module 403 is further configured to calculate a user value coefficient based on the second estimation data and the third estimation data; The estimation module 403 is further used to correct the first estimation data by using the user value coefficient to obtain a target estimated insurance user value.
[0059] In a possible implementation, the determination module 402 is further used to determine the corresponding target holiday based on the insurance time; and determine a target holiday insurance user value estimation model and a target non-holiday insurance user value estimation model based on the target holiday.
[0060] In one possible implementation, the estimation module 403 is also used to calculate a user value coefficient based on the second estimated data and the third estimated data through a first formula, wherein the first formula is: y=x1 / x2; wherein y is the user value coefficient, x1 is the second estimated data, and x2 is the third estimated data.
[0061] In one possible implementation, the estimation module 403 is also used to correct the first estimated data through the user value coefficient based on a second formula to obtain a target estimated insurance user value, and the second formula is: L=z×y; wherein L is the target estimated insurance user value, z is the first estimated data, and y is the user value coefficient.
[0062] In one possible embodiment, the insurance user value estimation device provided by the present invention also includes a model training module, which is used to obtain the first user insurance policy data during the target holiday and the second user insurance policy data before and after the target holiday; perform data feature processing on the first user insurance policy data and the second user insurance policy data to obtain the first sample user insurance policy data and the second sample user insurance policy data after data processing; divide the first sample user insurance policy data and the second sample user insurance policy data into a model training set and a model test set respectively; train the holiday insurance user value estimation model through the model training set in the first sample user insurance policy data, and train the non-holiday insurance user value estimation model through the model training set in the second sample user insurance policy data; test the holiday insurance user value estimation model after the initial training through the model test set in the first sample user insurance policy data to obtain a first test result, and test the non-holiday insurance user value estimation model after the initial training through the model test set in the second sample user insurance policy data to obtain a second test result; when the first test result and the second test result meet the preset conditions, it is determined that the training of the holiday insurance user value estimation model and the non-holiday insurance user value estimation model is completed, and the target insurance user value estimation model is obtained.
[0063] Figure 5 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 5As shown, the electronic device may include: a processor (processor) 810, a communication interface (Communications Interface) 820, a memory (memory) 830 and a communication bus 840, wherein the processor 810, the communication interface 820, and the memory 830 communicate with each other through the communication bus 840. The processor 810 can call the logic instructions in the memory 830 to execute the insurance user value estimation method, which includes: obtaining the target insurance policy data of the target user and determining the insurance time based on the target insurance policy data; determining the target insurance user value estimation model based on the insurance time, the target insurance user value estimation model includes a target holiday insurance user value estimation model and a target non-holiday insurance user value estimation model, and the target holiday insurance user value estimation model and the target non-holiday insurance user value estimation model are obtained by training the user insurance policy data during the target holiday period and the user insurance policy data before and after the target holiday; inputting the target insurance policy data into the initial insurance user value estimation model, the target holiday insurance user value estimation model and the target non-holiday insurance user value estimation model respectively, and obtaining the first estimation data, the second estimation data and the third estimation data output by the models respectively; calculating the user value coefficient based on the second estimation data and the third estimation data; correcting the first estimation data by the user value coefficient to obtain the target estimated insurance user value.
[0064] In addition, the logic instructions in the above-mentioned memory 830 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.
[0065] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the insurance user value estimation method provided by the above-mentioned methods, which includes: obtaining the target insurance policy data of the target user and determining the insurance time based on the target insurance policy data; determining a target insurance user value estimation model based on the insurance time, the target insurance user value estimation model including a target holiday insurance user value estimation model and a target non-holiday insurance user value estimation model, the target holiday insurance user The user value estimation model and the target non-holiday insurance user value estimation model are obtained by training the user insurance policy data during the target holidays and the user insurance policy data before and after the target holidays; the target insurance policy data are respectively input into the initial insurance user value estimation model, the target holiday insurance user value estimation model and the target non-holiday insurance user value estimation model to obtain the first estimated data, the second estimated data and the third estimated data output by the models respectively; the user value coefficient is calculated based on the second estimated data and the third estimated data; the first estimated data is corrected by the user value coefficient to obtain the target estimated insurance user value.
[0066] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented by a processor to execute the insurance user value estimation method provided by the above-mentioned methods when executed, the method comprising: obtaining target insurance policy data of a target user and determining an insurance time based on the target insurance policy data; determining a target insurance user value estimation model based on the insurance time, the target insurance user value estimation model comprising a target holiday insurance user value estimation model and a target non-holiday insurance user value estimation model, the target holiday insurance user value estimation model and the target non-holiday insurance user value estimation model being obtained by training with user insurance policy data during the target holiday period and user insurance policy data before and after the target holiday; inputting the target insurance policy data into the initial insurance user value estimation model, the target holiday insurance user value estimation model and the target non-holiday insurance user value estimation model, respectively, to obtain first estimation data, second estimation data and third estimation data outputted by the models respectively; calculating a user value coefficient based on the second estimation data and the third estimation data; correcting the first estimation data by the user value coefficient to obtain a target estimated insurance user value.
[0067] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.
[0068] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.
[0069] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A method for estimating insurance user value, characterized in that: include: Acquire target insurance policy data of a target user and determine the insurance time based on the target insurance policy data; Determine a target insurance user value estimation model based on the insurance time, the target insurance user value estimation model includes a target holiday insurance user value estimation model and a target non-holiday insurance user value estimation model, the target holiday insurance user value estimation model and the target non-holiday insurance user value estimation model are obtained by training the user insurance policy data during the target holiday period and the user insurance policy data before and after the target holiday; Inputting the target insurance policy data into the initial insurance user value estimation model, the target holiday insurance user value estimation model and the target non-holiday insurance user value estimation model respectively, and obtaining the first estimation data, the second estimation data and the third estimation data outputted by the models respectively; Calculate a user value coefficient based on the second estimated data and the third estimated data; The first estimated data is corrected by using the user value coefficient to obtain a target estimated insurance user value.
2. The method according to claim 1, characterized in that: The calculating the user value coefficient based on the second estimated data and the third estimated data includes: The user value coefficient is calculated by a first formula based on the second estimated data and the third estimated data, wherein the first formula is: y=x1 / x2; Among them, y is the user value coefficient, x1 is the second estimated data, and x2 is the third estimated data.
3. The method according to claim 2, characterized in that Correcting the first estimated data by using the user value coefficient to obtain a target estimated insurance user value includes: Based on a second formula, the first estimated data is corrected by the user value coefficient to obtain a target estimated insurance user value, wherein the second formula is: L=z×y; Among them, L is the target estimated insurance user value, z is the first estimated data, and y is the user value coefficient.
4. The method according to claim 1, characterized in that: The target insurance user value estimation model determined based on the insurance time includes: Determine the corresponding target holiday based on the insurance time; A target holiday insurance user value estimation model and a target non-holiday insurance user value estimation model are determined based on the target holidays.
5. The method according to claim 1, characterized in that The method further comprises: The target insurance user value estimation model is trained by the following steps: Obtaining the insurance policy data of the first user during the target holiday and the insurance policy data of the second user before and after the target holiday; Performing data feature processing on the first user insurance policy data and the second user insurance policy data to obtain first sample user insurance policy data and second sample user insurance policy data after data processing; Dividing the first sample user insurance policy data and the second sample user insurance policy data into a model training set and a model test set respectively; Training a holiday insurance user value estimation model using the model training set in the first sample user's insurance policy data, and training a non-holiday insurance user value estimation model using the model training set in the second sample user's insurance policy data; The holiday insurance user value estimation model after initial training is tested by the model test set in the insurance policy data of the first sample user to obtain a first test result, and the non-holiday insurance user value estimation model after initial training is tested by the model test set in the insurance policy data of the second sample user to obtain a second test result; When the first test result and the second test result meet the preset conditions, it is determined that the training of the holiday insurance user value estimation model and the non-holiday insurance user value estimation model is completed, and the target insurance user value estimation model is obtained.
6. The method according to claim 5, characterized in that The test results shall at least include the model AUC evaluation value, estimated accuracy and model recall rate; When the first test result and the second test result meet the preset conditions, it is determined that the training of the holiday insurance user value estimation model and the non-holiday insurance user value estimation model is completed, and the target insurance user value estimation model is obtained, including: Determine whether the AUC evaluation value, the estimated accuracy rate, and the model recall rate in the first test result and the second test result respectively reach corresponding evaluation thresholds; If so, it is determined that the holiday insurance user value estimation model and the non-holiday insurance user value estimation model are trained and the target insurance user value estimation model is obtained; If not, the model parameters are modified and the holiday insurance user value estimation model is re-trained using the model training set in the first sample user's insurance policy data, and the non-holiday insurance user value estimation model is trained using the model training set in the second sample user's insurance policy data.
7. An insurance user value estimation device, characterized in that: include: An acquisition module, used to acquire target insurance policy data of a target user and determine the insurance time based on the target insurance policy data; A determination module is used to determine a target insurance user value estimation model based on the insurance time, wherein the target insurance user value estimation model includes a target holiday insurance user value estimation model and a target non-holiday insurance user value estimation model, wherein the target holiday insurance user value estimation model and the target non-holiday insurance user value estimation model are obtained by training the user insurance policy data during the target holiday period and the user insurance policy data before and after the target holiday; An estimation module, used for inputting the target insurance policy data into the initial insurance user value estimation model, the target holiday insurance user value estimation model and the target non-holiday insurance user value estimation model, respectively, to obtain first estimation data, second estimation data and third estimation data outputted by the models respectively; The estimation module is further used to calculate a user value coefficient based on the second estimation data and the third estimation data; The estimation module is further used to correct the first estimation data by using the user value coefficient to obtain a target estimated insurance user value.
8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the insurance user value estimation method as described in any one of claims 1 to 6 is implemented.
9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the insurance user value estimation method according to any one of claims 1 to 6 is implemented.
10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the insurance user value estimation method according to any one of claims 1 to 6 is implemented.