CN114612132A - Client renewal prediction method based on machine learning and related equipment - Google Patents

Abstract

The present application provides a machine learning-based customer insurance renewal prediction method and related equipment. The method includes: acquiring renewal policy data for model training; extracting multi-dimensional feature data from the renewal policy data; The multi-dimensional feature data trains the preset machine learning model to obtain the renewal prediction model; obtains the historical policy data of the target customer; inputs the historical policy data into the renewal prediction model, and outputs the target customer's corresponding policy renewal Guaranteed forecast results. The method of the present application can quickly and stably predict the customer's insurance renewal willingness, which is helpful for marketers to develop business according to the renewal prediction result, provide differentiated services to customers, and improve the insurance renewal rate.

Description

Machine Learning-Based Customer Renewal Prediction Method and Related Equipment

technical field

The present application relates to the field of computer technology, and in particular, to a method and related equipment for predicting customer renewal based on machine learning.

Background technique

In recent years, with the development of the insurance business, the number of insurance policies sold through the Internet has also increased year by year. For insurance companies, the renewal operation of the policy will directly affect the premium scale. When the insurance product purchased by the insured customer expires, the insured customer will be expected to renew the insurance. During this process, the marketer cannot know the customer's renewal intention, so it is impossible to carry out differentiated and efficient business development.

At present, the customer's willingness to renew insurance is generally predicted based on the personal experience of the marketer. Due to the differences in the personal experience of different marketers, the accuracy of human judgment is not high and unstable. Therefore, there is an urgent need for a solution for customer renewal forecasting, which can effectively and accurately predict customer renewal intentions.

SUMMARY OF THE INVENTION

In view of this, the purpose of this application is to propose a method for predicting customer renewal based on machine learning and related equipment to solve the above problems.

Based on the above purpose, a first aspect of the present application provides a method for predicting customer renewal based on machine learning, including:

Obtain renewal policy data for model training;

extracting multi-dimensional feature data from the renewal policy data;

Train a preset machine learning model according to the multi-dimensional feature data to obtain a warranty renewal prediction model;

Obtain historical policy data of target customers;

Inputting the historical policy data into the insurance renewal prediction model, and outputting the renewal prediction result of the policy corresponding to the target customer.

Further, according to the multi-dimensional feature data training preset machine learning model, the obtained renewal prediction model includes:

preprocessing the multidimensional feature data,

Divide the preprocessed multidimensional feature data into training datasets, validation datasets and test datasets;

The machine learning model is trained by using the training data set, and the accuracy of the machine learning model is obtained through the validation set every time an iteration is completed; and

Optimize the hyperparameters of the machine learning model through a hyperparameter optimization algorithm;

Repeat the above operation until the predetermined number of iterations is met or the accuracy of the machine learning model meets the preset condition, and a trained warranty renewal prediction model is obtained;

The accuracy of the trained warranty renewal prediction model is verified by using the test set to obtain a warranty renewal prediction result.

Further, after the inputting the historical policy data into the insurance renewal prediction model, and outputting the renewal prediction result of the corresponding insurance policy of the target customer, the method further includes:

Determine the renewal forecast probability of the target customer according to the renewal forecast result;

In response to determining that the predicted renewal probability is greater than or equal to a preset renewal probability threshold, the target customer is marked as a renewal intention customer and displayed.

Further, the preprocessing includes data cleaning and data dimensionality reduction.

Further, the hyperparameter optimization algorithm includes at least one of the following: a grid search algorithm, a random search algorithm and/or a Bayesian optimization algorithm.

Further, the multi-dimensional feature data includes: basic customer information, historical insurance application record information, historical claim settlement record information, insurance policy information, and basic information on salespersons.

Further, the algorithms used by the machine learning model include: random forest algorithm, Xgboost algorithm or Wide&Deep algorithm.

Based on the same inventive concept, a second aspect of the present application provides a device for predicting customer renewal based on machine learning, including:

a first acquisition module, configured to acquire policy renewal data for model training;

a feature extraction module configured to extract multi-dimensional feature data from the renewal policy data;

A model training module, configured to train a preset machine learning model according to the multi-dimensional feature data to obtain a warranty renewal prediction model;

The second acquisition module is configured to acquire historical policy data of the target customer;

The insurance renewal prediction module is configured to input the historical insurance policy data into the insurance renewal prediction model, and output the insurance renewal prediction result of the corresponding insurance policy of the target customer.

Based on the same inventive concept, a third aspect of the present application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the first aspect when executing the program the method described.

Based on the same inventive concept, a fourth aspect of the present application provides a non-transitory computer-readable storage medium, where the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions are used to cause a computer to execute the first aspect. Methods.

It can be seen from the above that the machine learning-based customer insurance renewal prediction method and related equipment provided in this application are based on machine learning and use the renewed insurance policy to train the machine learning model to obtain a trained insurance renewal prediction model. , using the trained insurance renewal prediction model to quickly and stably predict the renewal intention of the predicted customers, which will help the marketers to expand their business according to the renewal prediction results, provide differentiated services to customers, and improve the renewal rate.

Description of drawings

In order to illustrate the technical solutions in the present application or related technologies more clearly, the following briefly introduces the accompanying drawings used in the description of the embodiments or related technologies. Obviously, the drawings in the following description are only for the present application. In the embodiments, for those of ordinary skill in the art, other drawings can also be obtained according to these drawings without any creative effort.

1 is a flowchart of a method for predicting customer renewal based on machine learning according to an embodiment of the application;

2 is a flowchart of a training method for a warranty renewal prediction model according to an embodiment of the application;

FIG. 3 is a flowchart of a method for displaying renewal prediction results according to an embodiment of the application;

4 is a schematic structural diagram of a device for predicting customer renewal based on machine learning according to an embodiment of the application;

FIG. 5 is a schematic structural diagram of an electronic device according to an embodiment of the present application.

Detailed ways

In order to make the objectives, technical solutions and advantages of the present application more clearly understood, the present application will be further described in detail below with reference to specific embodiments and accompanying drawings.

It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the embodiments of the present application shall have the usual meanings understood by those with ordinary skills in the field to which the present application belongs. "First", "second" and similar words used in the embodiments of the present application do not indicate any order, quantity or importance, but are only used to distinguish different components. "Comprises" or "comprising" and similar words mean that the elements or things appearing before the word encompass the elements or things recited after the word and their equivalents, but do not exclude other elements or things. Words like "connected" or "connected" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "Down", "Left", "Right", etc. are only used to represent the relative positional relationship, and when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

As described in the background art section, the prediction scheme of the renewal intention in the related art is still difficult to meet the needs. Generally, the customer's renewal intention is judged by the personal experience of the salesman. In the process of realizing the present application, the applicant found that the prediction scheme of the renewal willingness in the prior art has at least the following problems: there is uncertainty in judging the customer's renewal willingness only based on the personal experience of the marketer. Marketers with longer working hours have richer personal experience than those with shorter working hours. Therefore, marketers with longer working hours are more accurate in judging customers' willingness to renew their insurance than those with shorter working hours. member. In addition, when the number of customers who need to predict the renewal willingness is large, the accuracy of manually judging the customer's renewal willingness cannot be guaranteed, and the work efficiency is greatly reduced, which is not conducive to the salesperson's business development.

In view of this, the embodiment of the present application provides a method for predicting customer renewal based on machine learning, extracting characteristic data from historical renewal policies, and using the characteristic data to train a machine learning model to obtain a renewal forecasting model, using The trained renewal prediction model can quickly and stably predict the customer's renewal intention.

Hereinafter, the technical solutions of the present application will be described in detail through specific embodiments.

Referring to FIG. 1 , a method for predicting customer renewal based on machine learning provided by an embodiment of the present application specifically includes the following steps:

Step S101, acquiring renewal policy data for model training.

In this step, renewal of insurance refers to the act of applying to the insurer to extend the term of the insurance contract or to go through the insurance procedures again when the insurance contract is about to expire. When going through the renewal procedures, the insurer or the insured can appropriately increase or decrease the insurance amount or make other changes according to the current objective situation or needs. For insurance companies, the renewal operation of the policy has a greater impact on the premium scale. Reasonable control of the renewal status of the policy that is about to expire will help increase the renewal rate and continue to generate premiums for the company.

The types of insurance are mainly divided into two categories: property insurance and personal insurance. Among them, property insurance can include: enterprise property insurance, engineering insurance, auto insurance, liability insurance, ship insurance, cargo insurance, home property insurance, credit insurance , Guarantee insurance, agricultural insurance; life insurance categories can include: accident insurance, medical insurance, critical illness insurance, life insurance, children's education insurance, endowment insurance, annuity insurance, group insurance. Different types of insurance have different renewing intentions of customers. For example, truck drivers need to drive trucks all year round, so truck drivers are more willing to renew insurance for accident insurance; people with poor health are more inclined to renew insurance for critical illness and medical insurance. High willingness to renew.

When obtaining the renewal policy data, the renewal policy can be extracted from each type of insurance to improve the prediction accuracy of the subsequently trained machine learning model. The policy data includes at least: the type of insurance purchased, the name of the insurance purchased, the regular premium, the salesperson, the total premium, customer asset information, customer age information, and the number of claims for the customer’s purchased insurance, etc.

Step S102, extracting multi-dimensional feature data from the renewal policy data.

In this step, the amount of data in the renewal policy is relatively large, so it is necessary to select the policy data that is conducive to improving the accuracy of the renewal prediction. Specifically, you can select basic customer information, historical insurance record information, historical claim record information, policy information and salesperson. The basic information is used to form multi-dimensional feature data. It should be noted that multi-dimensional feature data can also be added or reset according to the actual situation, which is not specifically limited here.

Step S103: Train a preset machine learning model according to the multi-dimensional feature data to obtain a warranty renewal prediction model.

In this step, the algorithm used by the machine learning model may be random forest algorithm, Xgboost algorithm or Wide&Deep algorithm.

The random forest algorithm refers to inputting training samples into each decision tree, and for each decision tree, randomly and with replacement part of the training samples are extracted as the training set of the tree, and the final classification result depends on multiple trees. the majority of classification results.

The Xgboost algorithm is an integrated machine learning algorithm based on decision trees. It adopts the Gradient Boosting framework and can be used for classification or regression by inputting parameter data.

The Wide&Deep algorithm aims to make the trained model gain both memory and generalization capabilities.

Among the above three algorithms, the Xgboost algorithm is the best choice, and it has good results in terms of classification accuracy, model training and model running speed.

Step S104, acquiring historical policy data of the target customer.

In this step, the target customer is the policyholder whose insurance contract is about to expire, and the obtained historical policy data is also the policy data corresponding to the policyholder about to expire. Predicting people's willingness to renew insurance is conducive to increasing the scale of insurance premiums.

Step S105: Input the historical policy data into the insurance renewal prediction model, and output the renewal prediction result of the policy corresponding to the target customer.

It can be seen that the machine learning-based customer insurance renewal prediction method provided in this embodiment is based on machine learning, and uses the renewed insurance policy to train the machine learning model, so as to obtain a trained insurance renewal prediction model, and use the trained insurance renewal prediction model. The model can quickly and stably predict the renewal intention of customers to be predicted, which helps marketers to develop businesses according to the renewal prediction results, provide differentiated services to customers, and improve the renewal rate.

In some embodiments, with reference to FIG. 2 , for step S103 in the foregoing embodiment, it may further include the following steps:

Step S1031, preprocessing the multi-dimensional feature data.

In this step, the preprocessing includes data cleaning and data dimensionality reduction, and the processing result directly affects the effect of the subsequent model. Among them, data cleaning includes processing of abnormal values, processing of null values, processing of missing values and duplicate values. For duplicate values and outliers, direct deletion can be used; for outliers, direct deletion or value substitution can be used; for null values, value substitution can be used to fill in missing values.

When the amount of data to be cleaned is large, a fixed period of time can be set to divide the data. In a period of time, a fixed period of data can be cleaned to avoid real-time cleaning of the data, due to the increase of data. Significant increase in the number of data cleanings to avoid multiple repeated cleanings of the data.

The specific steps of data dimensionality reduction are: standardizing the multi-dimensional feature data, the mean value is 0, and the variance is 1; calculating the covariance matrix, the eigenvalues of the covariance matrix and the eigenvectors corresponding to the eigenvalues; The eigenvalues are sorted by size, the largest k among the eigenvalues are selected, and the corresponding eigenvectors are used as row vectors to form an eigenvector matrix; in the new space.

The multi-dimensional feature data is processed by the preprocessing method of data cleaning and data dimension reduction, which greatly reduces the amount of training data and improves the training time and prediction accuracy of the model. By reducing the dimension of the data, the original data information is generated without losing the original data information. The large amount of new feature data greatly shortens the calculation time and improves the prediction ability of the model.

In addition, the samples with unbalanced distribution can be processed as follows: use Random, SMOTE, bSMOTE(1&2), SVM SMOTE, ADASYN and other upsampling methods in the imbalanced-learn development package to upsample the samples with a low proportion; Use Random, Tomek links, NearMiss and other downsampling methods to downsample the samples with a high proportion. When the algorithm applied by the machine learning model is the Xgboost algorithm, there is a specific parameter scale_pos_weight in the Xgboost algorithm. Its function is to adjust the error function of the model according to the set ratio for unbalanced samples, and increase the learning rate of a few samples. In this way, the influence of the misjudgment of low proportion samples on the model error is enhanced, so as to achieve the effect of adjusting the proportion of samples.

Step S1032: Divide the preprocessed multi-dimensional feature data into a training data set, a verification data set and a test data set.

In this step, the ratio of training data set, verification data set and test data set can be selected from 7:1:2 or 8:1:1, or the training data set, verification data set and test data set can be divided according to the actual data volume. The proportion is adjusted, which is not specifically limited here.

Step S1033: Use the training data set to train the machine learning model, and each time an iteration is completed, obtain the accuracy of the machine learning model through the verification set, and use the hyperparameter optimization algorithm to adjust the machine learning model. to optimize the hyperparameters.

In this step, the hyperparameters of the machine learning model may be optimized through a grid search algorithm, a random search algorithm and/or a Bayesian optimization algorithm. Among them, the grid search algorithm is a hyperparameter search method, which is specifically: in a hyperparameter space, according to a fixed step size, traverse the parameters on all step points, select the best hyperparameter point, and then continue to use this hyperparameter point. In the neighborhood of the parameter point, reduce the step size and continue to traverse until the set minimum step size is reached, and the obtained hyperparameter point is the optimal hyperparameter point. The random search algorithm is a random combination of hyperparameters. Compared with the network search algorithm, due to its randomness, it has a certain probability to reduce the amount of calculation. Bayesian optimization is to calculate the posterior probability distribution of the first n points through Gaussian process regression, and obtain the expected mean and variance of each set of hyperparameters of the model at the value point, where the mean represents the hyperparameter corresponding to this point, The expected effect of the model, the larger the mean, the better the final effect of the model; the variance represents the uncertainty of the effect, and the larger the variance, the greater the uncertainty of the effect obtained at this point.

Step S1034: Step S1033 is repeatedly executed until a predetermined number of iterations is satisfied or the accuracy of the machine learning model meets a preset condition, and a trained warranty renewal prediction model is obtained.

In this step, the preset condition is the accuracy threshold. It should be noted that the number of iterations and the accuracy threshold can be set according to actual conditions, which are not specifically limited here.

Step S1035, using the test set to verify the accuracy of the trained warranty renewal prediction model to obtain a warranty renewal prediction result.

In this step, when the test set is used to input the warranty renewal prediction model, and the corresponding warranty renewal prediction result satisfies the preset prediction accuracy, it means that the performance of the warranty renewal prediction model is good.

In some embodiments, referring to FIG. 3 , for step S105 in the foregoing embodiment, the following steps may be included after it:

Step S301, determining the renewal prediction probability of the target customer according to the renewal prediction result.

Step S302, in response to determining that the predicted renewal probability is greater than or equal to a preset renewal probability threshold, the target customer is marked as a renewal intention customer and displayed.

In this embodiment, by setting the renewal probability threshold, a customer with a higher willingness to renew is selected, and the customer is displayed to the corresponding salesperson, so that the salesperson can preferentially formulate a business development strategy. In addition, multiple insurance renewal probability thresholds can also be set to divide the customer's insurance renewal willingness into multiple levels, which is conducive to further improving the business development efficiency of marketers.

It should be noted that, the methods in the embodiments of the present application may be executed by a single device, such as a computer or a server. The method in this embodiment can also be applied in a distributed scenario, and is completed by the cooperation of multiple devices. In the case of such a distributed scenario, one device among the multiple devices may only execute one or more steps in the methods of the embodiments of the present application, and the multiple devices will interact with each other to complete all the steps. method described.

It should be noted that some embodiments of the present application are described above. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in an order different from that in the above-described embodiments and still achieve desirable results. Additionally, the processes depicted in the figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Based on the same inventive concept, and corresponding to the method in any of the above embodiments, the present application also provides a device for predicting customer renewal based on machine learning.

Referring to Figure 4, the machine learning-based customer warranty renewal prediction device includes:

The first acquisition module is configured to acquire the renewal policy data for model training.

A feature extraction module configured to extract multi-dimensional feature data from the renewal policy data.

The model training module is configured to train a preset machine learning model according to the multi-dimensional feature data to obtain a warranty renewal prediction model.

The second acquiring module is configured to acquire historical policy data of the target customer.

The insurance renewal prediction module is configured to input the historical insurance policy data into the insurance renewal prediction model, and output the insurance renewal prediction result of the corresponding insurance policy of the target customer.

As an optional embodiment, the model training module is specifically configured to preprocess the multidimensional feature data;

Divide the preprocessed multidimensional feature data into training datasets, validation datasets and test datasets;

The machine learning model is trained by using the training data set, and each iteration is completed, the accuracy of the machine learning model is obtained through the verification set; and the hyperparameters of the machine learning model are adjusted by a hyperparameter optimization algorithm. Perform optimization; repeat the above operations until a predetermined number of iterations are met or the accuracy of the machine learning model meets a preset condition, and a trained warranty renewal prediction model is obtained; use the test set to predict the trained warranty renewal The accuracy of the model is verified, and the renewal prediction results are obtained.

As an optional embodiment, the apparatus further includes a forecast result display module (not shown in the figure), and the forecast result display module is configured to determine the renewal forecast probability of the target customer according to the renewal forecast result ; In response to determining that the predicted renewal probability is greater than or equal to a preset renewal probability threshold, the target customer is marked as a renewal intention customer and displayed.

As an optional embodiment, the preprocessing includes data cleaning and data dimensionality reduction.

As an optional embodiment, the hyperparameter optimization algorithm includes at least one of the following: a grid search algorithm, a random search algorithm and/or a Bayesian optimization algorithm.

As an optional embodiment, the multi-dimensional feature data includes: basic customer information, historical insurance application record information, historical claim settlement record information, insurance policy information, and basic information of marketers.

As an optional embodiment, the algorithm used by the machine learning model includes: random forest algorithm, Xgboost algorithm or Wide&Deep algorithm.

For the convenience of description, when describing the above device, the functions are divided into various modules and described respectively. Of course, when implementing the present application, the functions of each module may be implemented in one or more software and/or hardware.

The apparatus of the foregoing embodiment is used to implement the corresponding machine learning-based customer warranty renewal prediction method in any of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiment, which will not be repeated here.

Based on the same inventive concept, corresponding to any of the above-mentioned embodiments, the present application further provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, the processor When the program is executed, the method for predicting customer renewal based on machine learning described in any one of the above embodiments is implemented.

FIG. 5 shows a schematic diagram of a more specific hardware structure of an electronic device provided in this embodiment. The device may include: a processor 1010 , a memory 1020 , an input/output interface 1030 , a communication interface 1040 and a bus 1050 . The processor 1010 , the memory 1020 , the input/output interface 1030 and the communication interface 1040 realize the communication connection among each other within the device through the bus 1050 .

The processor 1010 may be implemented by a general-purpose CPU (Central Processing Unit, central processing unit), a microprocessor, an application specific integrated circuit (Application Specific Integrated Circuit, ASIC), or one or more integrated circuits, and is used to execute related program to implement the technical solutions provided by the embodiments of this specification.

The memory 1020 may be implemented in the form of a ROM (Read Only Memory, read only memory), a RAM (Random Access Memory, random access memory), a static storage device, a dynamic storage device, and the like. The memory 1020 may store an operating system and other application programs. When implementing the technical solutions provided by the embodiments of this specification through software or firmware, the relevant program codes are stored in the memory 1020 and invoked by the processor 1010 for execution.

The input/output interface 1030 is used to connect the input/output module to realize information input and output. The input/output/module can be configured in the device as a component (not shown in the figure), or can be externally connected to the device to provide corresponding functions. The input device may include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output device may include a display, a speaker, a vibrator, an indicator light, and the like.

The communication interface 1040 is used to connect a communication module (not shown in the figure), so as to realize the communication interaction between the device and other devices. The communication module may implement communication through wired means (eg, USB, network cable, etc.), or may implement communication through wireless means (eg, mobile network, WIFI, Bluetooth, etc.).

Bus 1050 includes a path to transfer information between the various components of the device (eg, processor 1010, memory 1020, input/output interface 1030, and communication interface 1040).

It should be noted that although the above-mentioned device only shows the processor 1010, the memory 1020, the input/output interface 1030, the communication interface 1040 and the bus 1050, in the specific implementation process, the device may also include necessary components for normal operation. other components. In addition, those skilled in the art can understand that, the above-mentioned device may only include components necessary to implement the solutions of the embodiments of the present specification, rather than all the components shown in the figures.

The electronic device in the above embodiment is used to implement the corresponding machine learning-based customer warranty renewal prediction method in any of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiment, which will not be repeated here.

Based on the same inventive concept and corresponding to any of the methods in the above embodiments, the present application further provides a non-transitory computer-readable storage medium, where the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions use in causing the computer to execute the method for predicting customer renewal based on machine learning as described in any of the above embodiments.

The computer readable medium of this embodiment includes both permanent and non-permanent, removable and non-removable media and can be implemented by any method or technology for information storage. Information may be computer readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), Flash Memory or other memory technology, Compact Disc Read Only Memory (CD-ROM), Digital Versatile Disc (DVD) or other optical storage, Magnetic tape cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission medium that can be used to store information that can be accessed by a computing device.

The computer instructions stored in the storage medium of the above embodiments are used to cause the computer to execute the method for predicting customer renewal based on machine learning as described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here. Repeat.

Those of ordinary skill in the art should understand that the discussion of any of the above embodiments is only exemplary, and is not intended to imply that the scope of the application (including the claims) is limited to these examples; under the idea of the application, the above embodiments or Technical features in different embodiments can also be combined, steps can be implemented in any order, and there are many other variations of different aspects of the embodiments of the present application as described above, which are not provided in detail for the sake of brevity.

In addition, to simplify description and discussion, and to not obscure the understanding of the embodiments of the present application, well-known power/power sources associated with integrated circuit (IC) chips and other components may or may not be shown in the provided figures. ground connection. Furthermore, devices may be shown in block diagram form in order to avoid obscuring the embodiments of the present application, and this also takes into account the fact that details regarding the implementation of these block diagram devices are highly dependent on the implementation of the embodiments of the present application platform (ie, these details should be well within the understanding of those skilled in the art). Where specific details (eg, circuits) are set forth to describe exemplary embodiments of the present application, it will be apparent to those skilled in the art that these specific details may be used without or with changes to the specific details The embodiments of the present application are implemented below. Accordingly, these descriptions are to be considered illustrative rather than restrictive.

Although the present application has been described in conjunction with specific embodiments thereof, many alternatives, modifications, and variations to these embodiments will be apparent to those of ordinary skill in the art from the foregoing description. For example, other memory architectures (eg, dynamic RAM (DRAM)) may use the discussed embodiments.

The embodiments of the present application are intended to cover all such alternatives, modifications and variations that fall within the broad scope of the appended claims. Therefore, any omission, modification, equivalent replacement, improvement, etc. made within the spirit and principle of the embodiments of the present application shall be included within the protection scope of the present application.

Claims (10)

1. A client renewal prediction method based on machine learning is characterized by comprising the following steps:

acquiring continuous insurance policy data for model training;

extracting multi-dimensional characteristic data from the renewal warranty data;

training a preset machine learning model according to the multi-dimensional feature data to obtain a continuous prediction model;

acquiring historical policy data of a target client;

and inputting the historical policy data into the renewal prediction model, and outputting a renewal prediction result of the policy corresponding to the target customer.

2. The prediction method according to claim 1, wherein the training of a preset machine learning model according to the multidimensional feature data to obtain a continuous prediction model comprises:

preprocessing the multi-dimensional feature data;

dividing the preprocessed multidimensional characteristic data into a training data set, a verification data set and a test data set;

training the machine learning model by using the training data set, and acquiring the accuracy of the machine learning model through the verification set every time iteration is completed; and

optimizing the hyper-parameters of the machine learning model through a hyper-parameter optimization algorithm;

repeatedly executing the operation until the preset iteration times are met or the accuracy of the machine learning model meets the preset condition, and obtaining a trained continuous prediction model;

and verifying the accuracy of the trained continuous prediction model by using the test set to obtain a continuous prediction result.

3. The forecasting method of claim 1, wherein the step of inputting the historical policy data into the renewal prediction model and outputting the renewal prediction result for the policy corresponding to the target customer further comprises:

determining the renewal prediction probability of the target client according to the renewal prediction result;

in response to determining that the renewal prediction probability is greater than or equal to a preset renewal probability threshold, marking the target customer as an renewal intent customer and presenting.

4. The prediction method of claim 2, wherein the preprocessing comprises data cleansing and data dimensionality reduction.

5. The prediction method of claim 2, wherein the hyper-parametric optimization algorithm comprises at least one of: a grid search algorithm, a random search algorithm, and/or a bayesian optimization algorithm.

6. The prediction method according to any one of claims 1 to 5, wherein the multi-dimensional feature data comprises: the system comprises client basic information, historical insurance application record information, historical claim settlement record information, insurance policy information and marketer basic information.

7. The prediction method according to any one of claims 1 to 5, wherein the machine learning model uses an algorithm comprising: a random forest algorithm, an Xgboost algorithm, or a Wide & Deep algorithm.

8. A client renewal prediction device based on machine learning, comprising:

a first acquisition module configured to acquire renewal sheet data for model training;

a feature extraction module configured to extract multi-dimensional feature data from the renewal warranty data;

the model training module is configured to train a preset machine learning model according to the multi-dimensional feature data to obtain a continuous prediction model;

a second obtaining module configured to obtain historical policy data of the target customer;

and the renewal prediction module is configured to input the historical policy data into the renewal prediction model and output a renewal prediction result of the policy corresponding to the target customer.

9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the method according to any of claims 1 to 7 when executing the program.

10. A non-transitory computer readable storage medium storing computer instructions for causing a computer to perform the method of any one of claims 1 to 7.

Images