Insurance industry potential customer mining method and device, and storage medium

Abstract

The application discloses an insurance industry potential customer mining method, and is applied to the technical field of data mining. The method provided by the application comprises the following steps: acquiring a customer information set of a target customer to be analyzed from a preset customer acquisition channel; performing analysis on the customer information to obtain a user portrait label of the target customer; processing the user portrait label of the target customer according to a first screening rule to obtain target customer data to be trained; performing machine learning training on the target customer data to be trained by using a Spark computing platform to obtain a potential customer prediction model; inputting the target customer data to be trained into the customer prediction model to obtain first potential customer data; processing the user portrait label of the target customer according to a second screening rule to obtain second potential customer data; pushing the first potential customer data and the second potential customer data to an operation platform; and tracking changes of the first potential customer data and the second potential customer data.

Description

Technical Field

[0001] This application relates to the field of data mining technology, and in particular to a method, apparatus, computer equipment, and storage medium for mining potential customers in the insurance industry. Background Technology

[0002] As the insurance business continues to grow, customer acquisition costs are rising, and acquiring new customers is becoming increasingly difficult. Therefore, it is necessary to deeply explore the value of existing customers and maintain their loyalty. This necessitates potential customer acquisition and cross-operation with existing customers across various sub-sectors of the insurance industry.

[0003] Traditional customer acquisition and cross-operation in the insurance industry primarily rely on the industry experience of insurance agents in various regions or coarse-grained recommendations from relevant customer management systems. Cross-operation also heavily depends on the agents' industry experience and understanding of basic customer profiles. This traditional, experience-dependent approach to customer acquisition is not conducive to new agents, and it fails to further utilize other customer data or provide refined, prioritized, and tiered customer management. Summary of the Invention

[0004] This application provides a method, apparatus, computer equipment, and storage medium for mining potential customers in the insurance industry, in order to solve the problem that traditional potential customer mining techniques in the insurance industry rely heavily on personal experience and lack refined operation.

[0005] The first aspect of this application provides a method for identifying potential customers in the insurance industry, including:

[0006] Obtain a set of customer information for the target customers to be analyzed, wherein the customer information is obtained from preset customer acquisition channels according to preset customer screening rules;

[0007] Obtain user profile tags for the target customer, wherein the user profile tags are obtained by analyzing the customer information;

[0008] Obtain target customer data to be trained, wherein the target customer data to be trained is obtained by processing the user profile tags of the target customers according to the first screening rule;

[0009] The target customer data to be trained is input into the Spark computing platform to obtain the first potential customer data, wherein the first potential customer data is the output of the potential customer prediction model, and the potential customer prediction model is obtained by the Spark computing platform using the target customer data to be trained for machine learning training.

[0010] Obtain second potential customer data, wherein the second potential customer data is obtained by processing the user profile tags of the target customer according to the second screening rule;

[0011] The first potential customer data and the second potential customer data are pushed to the corresponding operation platform, and the changes in the first potential customer data and the second potential customer data are tracked and displayed using a preset reporting system.

[0012] A second aspect of this application provides a potential customer mining device for the insurance industry, comprising:

[0013] The customer information collection module is used to obtain a collection of customer information of the target customers to be analyzed, wherein the customer information is obtained from preset customer acquisition channels according to preset customer screening rules;

[0014] The user profile tag module is used to obtain user profile tags for the target customer, wherein the user profile tags are obtained by analyzing the customer information;

[0015] The target customer data module is used to acquire target customer data to be trained, wherein the target customer data to be trained is obtained by processing the user profile tags of the target customers according to the first screening condition;

[0016] The first potential customer data module is used to input the target customer data to be trained into the Spark computing platform to obtain the first potential customer data. The first potential customer data is output by the potential customer prediction model, which is obtained by the Spark computing platform using the target customer data to be trained for machine learning training.

[0017] The second potential customer data module is used to acquire second potential customer data, wherein the second potential customer data is obtained by processing the user profile tags of the target customer according to the second screening conditions;

[0018] The reporting system tracking module is used to push the first potential customer data and the second potential customer data to the corresponding operation platform, and use a preset reporting system to track and display the changes in the first potential customer data and the second potential customer data.

[0019] A third aspect of this application provides a computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described insurance industry potential customer mining method.

[0020] A fourth aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described method for identifying potential customers in the insurance industry.

[0021] The aforementioned method, apparatus, computer equipment, and storage medium for potential customer mining in the insurance industry involve: acquiring a set of customer information for target customers to be analyzed from preset customer acquisition channels; analyzing the customer information to obtain user profile tags for the target customers; processing the user profile tags of the target customers according to a first screening rule to obtain target customer data to be trained; using the target customer data to be trained for machine learning training on a Spark computing platform to obtain a potential customer prediction model; inputting the target customer data to be trained into the customer prediction model to obtain first potential customer data; processing the user profile tags of the target customers according to a second screening rule to obtain second potential customer data; pushing the first and second potential customer data to an operations platform; and using a preset reporting system to track changes in the first and second potential customer data. This not only makes potential customer mining in the insurance industry independent of personal experience but also fully utilizes existing customer user data and achieves refined operations through multi-layered mining of potential customer data in different ways and tracking the actual effects of potential customer data. Attached Figure Description

[0022] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a schematic diagram of an application environment for a method for identifying potential customers in the insurance industry, as described in one embodiment of this application.

[0024] Figure 2 This is a flowchart of a method for identifying potential customers in the insurance industry according to one embodiment of this application;

[0025] Figure 3 This is a schematic diagram of the structure of a potential customer discovery device for the insurance industry in one embodiment of this application;

[0026] Figure 4 This is a schematic diagram of a computer device according to one embodiment of this application. Detailed Implementation

[0027] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0028] The method for identifying potential customers in the insurance industry provided in this application can be applied to, for example... Figure 1 In this application environment, the computer equipment can be, but is not limited to, various personal computers and laptops. The computer equipment can also be a server, which can be a standalone server or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms. This is understandable. Figure 1 The number of computer devices shown is merely illustrative and can be expanded in any number according to actual needs.

[0029] In one embodiment, such as Figure 2 As shown, a method for identifying potential customers in the insurance industry is provided, and this method is applied to... Figure 1 The following steps, S101 to S106, are used as an example of computer equipment in the example:

[0030] S101. Obtain a set of customer information of the target customers to be analyzed, wherein the customer information is obtained from a preset customer acquisition channel according to preset customer screening rules.

[0031] The preset customer acquisition channels can be any information system or platform where the insurance company stores customer information, such as a customer relationship management system or platform. They can also be third-party platforms that interface with the insurance company, such as internet insurance sales platforms, or subsystems or platforms for different types of insurance, such as life insurance subsystems or platforms, property insurance subsystems or platforms, and pension insurance subsystems or platforms. The preset customer screening conditions include, but are not limited to: restrictions on customer source channels, restrictions on the customer's region, restrictions on the customer's age range, restrictions on the customer's asset registration, and restrictions on the customer's activity level. Furthermore, the preset customer screening conditions are obtained by insurance professionals through fundamental analysis of the insurance business scenario, and are also pre-defined with coarse-grained restrictions based on business needs after obtaining potential customers. For example, if only a portion of the target insurance products are promoted to the target potential customers obtained using the potential customer mining method provided in this embodiment, then coarse-grained restrictions are imposed based on the demographic characteristics covered by the target insurance products. For instance, if promoting an accident insurance policy for primary and secondary school students to potential customers, the coarse-grained restriction before potential customer mining is whether the customer has immediate family members currently in primary or secondary school.

[0032] S102. Obtain the user profile tags of the target customer, wherein the user profile tags are obtained by analyzing the customer information.

[0033] Since the customer acquisition channels differ, the acquired customer information possesses various dimensions. These different dimensions of customer information are then used to create a comprehensive customer profile. The user profile tags include, but are not limited to: historical policy data, historical insurance enrollment information, personal information, historical claims information, behavioral activity information, personal asset information, personal interests and preferences, family information, types of insurance products held, number of customer service calls, call duration, and historical complaints. It should be noted that, because the customer information involves a significant amount of personal information, de-identification processing is required during the processing of this customer information. This de-identification process is not a core step in this embodiment and will not be elaborated upon here. Specifically, in this embodiment, the acquired customer information from different dimensions is first processed and organized, and then feature derivation engineering is used to create a comprehensive customer profile, resulting in the user profile tags for the target customer.

[0034] S103. Obtain target customer data to be trained, wherein the target customer data to be trained is obtained by processing the user profile tags of the target customers according to the first screening rule.

[0035] Specifically, processing the user profile tags of the target customer according to the first screening rule includes: First, calculating the tag saturation of the user profile tags and removing user profile tags whose tag saturation is not within a preset tag saturation range. Second, calculating the information value of the user profile tags and removing user profile tags whose information value is not within a preset information value range, that is, predicting the predictive ability of the user profile tags and selecting user profile tags with high information contribution and better ability to distinguish between potential customers and non-potential customers. Third, inputting the user profile tags into a preset feature importance screening model and outputting user profile tags whose feature importance scores are within a preset importance score range, for example, inputting the user profile tags into a pre-trained random forest model and outputting the importance score results of the user profile tags, and then sorting and screening the user profile tags according to the importance scores. Fourthly, the similarity and relevance values ​​of the user profile tags are calculated. User profile tags with relevance values ​​within a preset range and those with the highest similarity to the target user profile tag are retained. For example, the Pearson product-moment correlation coefficient is used to calculate the relevance matrix of the user profile tags. Only user profile tags with an absolute relevance value greater than 0.5 to the target user profile tag are retained. Furthermore, for a group of user profile tags with high similarity, only those with a high relevance to the target user profile tag are retained, for example, only those with a relevance within a preset range. Fifthly, the stability score of the user profile tags is calculated according to preset tag stability calculation rules. Due to the special nature of cross-business scenarios in the insurance industry, the user profile tags partially rely on information output from external third-party platforms. Therefore, it is necessary to evaluate the stability of the information output from these third-party platforms, i.e., the stability of the user profile tags, and then remove user profile tags with low stability scores, or use them as auxiliary tags for subsequent customer segmentation. For example, the timeliness and frequency of information output from the third-party platform can be used as information elements in calculating the stability score. More specific stability score calculations may vary depending on the specific business scenario and will not be elaborated here. Processing the user profile tags according to the first screening rule can merge user profile tags that do not meet the threshold requirements into one category to reduce the unevenness of feature distribution. Furthermore, for continuous user profile tags, they are filled according to corresponding business needs, and clustering binning is used to bin the continuous user profile tags to obtain categorical feature user profile tags, making the user profile tags more conducive to subsequent potential customer prediction model training.

[0036] S104. Input the target customer data to be trained into the Spark computing platform to obtain the first potential customer data, wherein the first potential customer data is the output of the potential customer prediction model, and the potential customer prediction model is obtained by the Spark computing platform using the target customer data to be trained for machine learning training.

[0037] The Spark computing platform is built upon the Apache Spark project, a distributed open-source processing system for big data workloads. Utilizing in-memory caching and optimized query execution, it enables rapid analysis and querying of data of any scale. Apache Spark provides development APIs in Java, Scala, Python, and R, and supports code reuse across multiple workloads, such as batch processing, interactive queries, real-time analytics, machine learning, and graph processing. Apache Spark comprises five core components: Spark Core, Spark SQL, Spark Streaming, Spark MLlib, and Spark GraphX. Furthermore, the Spark SQL component provides a distributed query engine for low-latency interactive queries, which can be used to retrieve customer information. Spark Streaming can extract data from batch processing for streaming analysis, and can be used to obtain large volumes of real-time customer information data from third-party platforms. Spark MLlib is an algorithm library required for machine learning on large-scale data, and can be used to train the potential customer model for machine learning. Spark GraphX ​​is a distributed graph processing framework that provides ETL, exploratory analytics, and iterative graph computation. It can be used to process client information with graphical data structures.

[0038] Before inputting the target customer data to be trained into the Spark computing platform to obtain the first potential customer data, the method further includes: processing the target customer data to be trained using an undersampling method to ensure that the ratio of the number of positive examples to the number of negative examples in the target customer data to be trained is within a preset range. The positive examples are the target customer data labeled as real potential customers, and the negative examples are the target customer data labeled as non-real potential customers. The LightBGM ensemble tree model is selected as the potential customer prediction model after comprehensively evaluating the training effects of various machine learning methods, including the LightBGM ensemble tree model, Random Forest, Logistic Regression (LR), Gradient Boosting Decision Tree (GBDT) + Logistic Regression model prediction, and neural networks, using machine learning model evaluation metrics such as AUC (Area Under The Curve) and ROC (Receiver Operating Characteristic Curve).

[0039] The process of inputting the target customer data to be trained into the Spark computing platform to obtain the first potential customer data includes: First, merging the LightBGM ensemble tree model into the SparkML pipeline of the Spark computing platform. Then, training the LightBGM ensemble tree model using the target customer data to obtain the trained potential customer prediction model. Finally, inputting the user profile tags of the target customers into the potential customer prediction model, and outputting the first potential customer data. Further, the LightBGM ensemble tree model outputs a potential customer prediction score for the target customers, sorts the target customers according to the potential customer prediction score from largest to smallest, and classifies the target customers into target potential customers and non-target potential customers according to a preset potential customer prediction score range. Finally, the prediction result and classification result for the target customers are used as the first potential customer data.

[0040] S105. Obtain second potential customer data, wherein the second potential customer data is obtained by processing the user profile tags of the target customer according to the second screening rule.

[0041] Specifically, firstly, the user profile tags of the target customers corresponding to the positive example data of the target customer data to be trained are obtained as tags to be calculated. Then, the unprocessed tag combinations of the unprocessed tag combinations are received, and the conversion rate of the target customers corresponding to the unprocessed tag combinations is calculated. Finally, the customer information of the target customers whose conversion rates are within a preset conversion rate range is obtained as the second potential customer data. Processing the user profile tags of the target customers according to the second screening rules can further utilize some customer information that was not used for training the potential customer model, increasing the success rate of mining potential customers. Furthermore, different sales strategies are associated with the corresponding second potential customer data according to different combinations of user profile tags, and the results of high and low conversion rates are mapped to the sales strategies. When the target customer matches multiple user profile tags or combinations of user profile tags, only the sales strategy with the highest conversion rate is associated. The setting of the sales strategies reduces the technical difficulty of operation and promotion between new insurance agents or insurance agents in different insurance sub-fields.

[0042] S106. Push the first potential customer data and the second potential customer data to the corresponding operation platform, and use the preset reporting system to track and display the changes in the first potential customer data and the second potential customer data.

[0043] Furthermore, the first potential customer data and the second potential customer data are fused to obtain target recommended potential customer data. This target potential customer data includes the potential value of each potential customer and associated sales strategies. Further, this target potential customer data is pushed to various insurance business operation platforms to help insurance agents quickly understand the needs of potential customers and achieve refined operations.

[0044] The process of tracking and displaying changes in the first and second potential customer data using a preset reporting system includes: First, tracking the first and second potential customer data using a preset FineBI reporting system. For example, using the FineBI expense reimbursement system to monitor the first and second potential customer data on a T+1 dimension. The monitored data includes the customer data conversion rate, i.e., the ratio of potential customer information corresponding to the first and second potential customer data that ultimately converts into actual customers generating new consumption behavior. Then, transmitting the monitoring data corresponding to the first potential customer data in the FineBI reporting system to the Spark computing platform to optimize the potential customer prediction model. Finally, using the monitoring data corresponding to the second potential customer data in the FineBI reporting system to optimize the second filtering rule.

[0045] This embodiment proposes a method for identifying potential customers in the insurance industry. The method involves: acquiring a set of customer information for target customers from pre-defined customer acquisition channels; analyzing the customer information to obtain user profile tags for the target customers; processing the user profile tags according to a first filtering rule to obtain target customer data for training; using the target customer data for training on a Spark computing platform to obtain a potential customer prediction model; inputting the target customer data for training into the customer prediction model to obtain first potential customer data; processing the user profile tags according to a second filtering rule to obtain second potential customer data; and pushing the first and second potential customer data to an operations platform, using a pre-defined reporting system to track changes in the first and second potential customer data. This method not only makes potential customer identification in the insurance industry independent of personal experience but also fully utilizes existing customer user data and achieves refined operations through multi-layered potential customer data mining and tracking.

[0046] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0047] In one embodiment, a potential customer mining device 100 for the insurance industry is provided, which corresponds one-to-one with the potential customer mining method for the insurance industry described in the above embodiments. For example... Figure 3 As shown, the insurance industry potential customer mining device 100 includes a customer information collection module 11, a user profile tagging module 12, a target customer data module 13, a first potential customer data module 14, a second potential customer data module 15, and a reporting system tracking module 16. Detailed descriptions of each functional module are as follows:

[0048] The customer information collection module 11 is used to obtain a collection of customer information of the target customers to be analyzed, wherein the customer information is obtained from a preset customer acquisition channel according to a preset customer screening rule;

[0049] User profile tag module 12 is used to obtain user profile tags of the target customer, wherein the user profile tags are obtained by analyzing the customer information;

[0050] The target customer data module 13 is used to acquire target customer data to be trained, wherein the target customer data to be trained is obtained by processing the user profile tags of the target customer according to the first screening condition;

[0051] The first potential customer data module 14 is used to input the target customer data to be trained into the Spark computing platform to obtain the first potential customer data, wherein the first potential customer data is the output of the potential customer prediction model, and the potential customer prediction model is obtained by the Spark computing platform using the target customer data to be trained for machine learning training.

[0052] The second potential customer data module 15 is used to obtain second potential customer data, wherein the second potential customer data is obtained by processing the user profile tags of the target customer according to the second screening conditions;

[0053] The reporting system tracking module 16 is used to push the first potential customer data and the second potential customer data to the corresponding operation platform, and use a preset reporting system to track and display the changes in the first potential customer data and the second potential customer data.

[0054] Furthermore, the target customer data module 13 also includes:

[0055] The tag saturation submodule is used to calculate the tag saturation of the user profile tags and remove user profile tags whose tag saturation is not within a preset tag saturation range;

[0056] The tag information quantum module is used to calculate the information value of the user profile tag and remove the user profile tag whose information value is not within the preset information value range;

[0057] The importance scoring submodule is used to input the user profile tags into a preset feature importance filtering model and output the user profile tags whose feature importance scores are within the preset importance score range.

[0058] The similarity and relevance submodule is used to calculate the similarity value and relevance value of the user profile tags, and retain the user profile tags whose relevance value is within a preset relevance value range and the user profile tags with the highest similarity to the target user profile tags.

[0059] Furthermore, the first potential customer data module 14 also includes:

[0060] The undersampling processing submodule is used to process the target customer data to be trained using an undersampling method, so that the ratio of the number of positive examples to the number of negative examples in the target customer data to be trained is within a preset range.

[0061] The ensemble tree model submodule is used to merge the LightBGM ensemble tree model into the SparkML pipeline of the Spark computing platform.

[0062] The prediction model training submodule is used to train the LightBGM ensemble tree model using the target customer data to be trained, so as to obtain the trained potential customer prediction model.

[0063] The prediction model application submodule is used to input the user profile tags of the target customer into the potential customer prediction model and output the first potential customer data.

[0064] Furthermore, the second potential customer data module 15 also includes:

[0065] The target customer data acquisition submodule is used to acquire the user profile labels of the target customers corresponding to the positive example data of the target customer data to be trained as labels to be calculated.

[0066] The conversion rate calculation submodule is used to receive the unprocessed tag combination of the unprocessed tag combination and calculate the conversion rate of the target customer corresponding to the unprocessed tag combination;

[0067] The conversion rate filtering submodule is used to obtain customer information of the target customers whose conversion rate is within a preset conversion rate range as the second potential customer data.

[0068] Furthermore, the reporting system tracking module 16 also includes:

[0069] The FineBI reporting system submodule is used to track the first potential customer data and the second potential customer data using a preset FineBI reporting system.

[0070] The first data optimization submodule is used to transmit the monitoring data corresponding to the first potential customer data in the FineBI reporting system to the Spark computing platform to optimize the potential customer prediction model.

[0071] The second data optimization submodule is used to optimize the second filtering rule using the monitoring data corresponding to the second potential customer data in the FineBI reporting system.

[0072] The terms "first" and "second" in the above-mentioned modules / units are only used to distinguish different modules / units and are not intended to specify which module / unit has a higher priority or any other limiting meaning. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or modules is not necessarily limited to those steps or modules explicitly listed, but may include other steps or modules not explicitly listed or inherent to these processes, methods, products, or devices. The module divisions appearing in this application are merely logical divisions; in actual applications, different division methods may be used.

[0073] Specific limitations regarding the potential customer mining device for the insurance industry can be found in the limitations of the potential customer mining method for the insurance industry mentioned above, and will not be repeated here. Each module in the aforementioned potential customer mining device for the insurance industry can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0074] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 4 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database stores data involved in potential customer mining methods in the insurance industry. The network interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a potential customer mining method for the insurance industry.

[0075] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the insurance industry potential customer mining method described in the above embodiments, for example... Figure 2 The steps S101 to S106 shown, as well as other extensions and related steps of the method, are examples. Alternatively, when the processor executes the computer program, it implements the functions of each module / unit of the insurance industry potential customer mining device in the above embodiments, for example... Figure 3 The functions of modules 11 to 12 are shown. To avoid repetition, they will not be described again here.

[0076] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the computer device, connecting various parts of the computer device via various interfaces and lines.

[0077] The memory can be used to store the computer programs and / or modules. The processor implements various functions of the computer device by running or executing the computer programs and / or modules stored in the memory and by calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, applications required for at least one function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created based on the use of the mobile phone (such as audio data, video data, etc.).

[0078] The memory can be integrated into the processor or it can be set up separately from the processor.

[0079] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When executed by a processor, the computer program implements the steps of the insurance industry potential customer mining method described in the above embodiments, for example... Figure 2 The steps S101 to S106 shown, as well as other extensions and related steps of the method, are examples. Alternatively, when a computer program is executed by a processor, it implements the functions of each module / unit of the insurance industry potential customer mining device in the above embodiments, for example... Figure 3 The functions of modules 11 to 16 are shown. To avoid repetition, they will not be described again here.

[0080] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0081] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0082] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for identifying potential customers in the insurance industry, characterized in that, include: Obtain a set of customer information for the target customers to be analyzed, wherein the customer information is obtained from preset customer acquisition channels according to preset customer screening rules; The user profile tags for the target customer are obtained, wherein the user profile tags are obtained by analyzing the customer information, and the user profile tags include: historical policy data, historical insurance information, personal information, historical claims information, behavioral activity information, personal asset information, personal interest and preference information, family information, types of insurance products held, number of customer service calls, customer service call duration, and historical complaints; the obtained customer information from different dimensions is first processed and organized, and then a feature derivation engineering method is used to create a comprehensive profile of the customer to obtain the user profile tags for the target customer; Obtain target customer data to be trained, wherein the target customer data to be trained is obtained by processing the user profile tags of the target customers according to the first screening rule; The target customer data to be trained is input into the Spark computing platform to obtain the first potential customer data, wherein the first potential customer data is the output of the potential customer prediction model, and the potential customer prediction model is obtained by the Spark computing platform using the target customer data to be trained for machine learning training. Obtain second potential customer data, wherein the second potential customer data is obtained by processing the user profile tags of the target customer according to the second screening rule; The first potential customer data and the second potential customer data are pushed to the corresponding operation platform, and the changes in the first potential customer data and the second potential customer data are tracked and displayed using a preset reporting system; The step of processing the user profile tags of the target customer according to the first filtering rule includes: calculating the tag saturation of the user profile tags, and removing the user profile tags whose tag saturation is not within a preset tag saturation range; calculating the information value of the user profile tags, and removing the user profile tags whose information value is not within a preset information value range; inputting the user profile tags into a preset feature importance filtering model, and outputting the user profile tags whose feature importance scores are within a preset importance score range; calculating the similarity value and relevance value of the user profile tags, and retaining the user profile tags whose relevance values ​​are within a preset relevance value range and the user profile tags with the highest similarity to the target user profile tags; The step of inputting the target customer data to be trained into the Spark computing platform to obtain the first potential customer data includes: merging the LightBGM ensemble tree model into the SparkML pipeline of the Spark computing platform; training the LightBGM ensemble tree model using the target customer data to be trained to obtain the trained potential customer prediction model; inputting the user profile tags of the target customers into the potential customer prediction model, and outputting the first potential customer data.

2. The method for identifying potential customers in the insurance industry according to claim 1, characterized in that, Before inputting the target customer data to be trained into the Spark computing platform to obtain the first potential customer data, the process also includes: The target customer data to be trained is processed using an undersampling method so that the ratio of the number of positive examples to the number of negative examples in the target customer data to be trained is within a preset ratio range.

3. The method for identifying potential customers in the insurance industry according to claim 1, characterized in that, The process of processing the user profile tags of the target customer according to the second screening rule includes: Obtain the user profile labels of the target customers corresponding to the positive example data of the target customer data to be trained as the labels to be calculated; Receive the combination of tags to be processed for the tags to be calculated, and calculate the conversion rate of the target customer corresponding to the combination of tags to be processed; The customer information of the target customers whose conversion rate is within a preset conversion rate range is obtained as the second potential customer data.

4. The method for identifying potential customers in the insurance industry according to claim 1, characterized in that, The method of using a preset reporting system to track and display changes in the first potential customer data and the second potential customer data includes: Use the preset FineBI reporting system to track the first potential customer data and the second potential customer data; Transmit the monitoring data corresponding to the first potential customer data in the FineBI reporting system to the Spark computing platform to optimize the potential customer prediction model; The second filtering rule is optimized using the monitoring data corresponding to the second potential customer data in the FineBI reporting system.

5. A device for identifying potential customers in the insurance industry, characterized in that, include: The customer information collection module is used to obtain a collection of customer information of the target customers to be analyzed, wherein the customer information is obtained from preset customer acquisition channels according to preset customer screening rules; The user profile tagging module is used to obtain user profile tags for the target customer. These user profile tags are obtained through analysis of the customer information and include: historical policy data, historical insurance enrollment information, personal information, historical claims information, behavioral activity information, personal asset information, personal interest and preference information, family information, types of insurance products held, number of customer service calls, customer service call duration, and historical complaints. The module first processes and organizes the obtained customer information from different dimensions, and then uses feature derivation engineering to create a comprehensive customer profile, resulting in the user profile tags for the target customer. The target customer data module is used to acquire target customer data to be trained, wherein the target customer data to be trained is obtained by processing the user profile tags of the target customers according to the first screening condition; The first potential customer data module is used to input the target customer data to be trained into the Spark computing platform to obtain the first potential customer data. The first potential customer data is output by the potential customer prediction model, which is obtained by the Spark computing platform using the target customer data to be trained for machine learning training. The second potential customer data module is used to acquire second potential customer data, wherein the second potential customer data is obtained by processing the user profile tags of the target customer according to the second screening conditions; The reporting system tracking module is used to push the first potential customer data and the second potential customer data to the corresponding operation platform, and use a preset reporting system to track and display the changes in the first potential customer data and the second potential customer data; The target customer data module further includes: a tag saturation submodule, used to calculate the tag saturation of the user profile tags and remove user profile tags whose tag saturation is not within a preset tag saturation range; a tag information quantum module, used to calculate the information content value of the user profile tags and remove user profile tags whose information content value is not within a preset information content value range; an importance scoring submodule, used to input the user profile tags into a preset feature importance filtering model and output user profile tags whose feature importance score is within a preset importance score range; and a similarity and relevance submodule, used to calculate the similarity value and relevance value of the user profile tags and retain user profile tags whose relevance value is within a preset relevance value range and the user profile tags with the highest similarity to the target user profile tags. The first potential customer data module further includes: an ensemble tree model submodule, used to merge the LightBGM ensemble tree model into the SparkML pipeline of the Spark computing platform; a prediction model training submodule, used to train the LightBGM ensemble tree model using the target customer data to be trained, to obtain the trained potential customer prediction model; and a prediction model application submodule, used to input the user profile tags of the target customer into the potential customer prediction model, and output the first potential customer data.

6. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the insurance industry potential customer mining method as described in any one of claims 1 to 5.

7. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the insurance industry potential customer mining method as described in any one of claims 1 to 5.

Images