Method and device for identifying purchase intention of insurance product by user

Abstract

The invention relates to the technical field of insurance marketing recommendation, in particular to a method and device for identifying the purchase intention of a user for an insurance product, and the method comprises the steps: collecting an identity label of the user and at least one of corresponding historical insurance information, insurance policy service information, complaint and consultation information and customer portrait information; based on a preset customer sign system, generating a customer data set according to the identity label of the user and at least one of corresponding historical insurance information, insurance policy service information, complaint and consultation information and customer portrait information; constructing a customer classification model by using the customer data set to determine main feature factors influencing the insurance purchasing intention of the customer; and constructing a clustering model to output the actual insurance product purchase intention of the target customer by using the clustering model. According to the method, customer behaviors can be deeply analyzed for customer purchase intention mining, products and corresponding services are accurately recommended according to customer demands, the marketing period is shortened, the conversion rate is increased, and the customer satisfaction is improved.

Description

Technical Field

[0001] The present application relates to the technical field of insurance marketing recommendation, and in particular to a method and device for identifying a user's intention to purchase an insurance product. Background Art

[0002] In the relevant technologies, in the health insurance marketing and customer development stage, sales managers usually roughly screen the target customer group based on their previous marketing experience, and then send clues to front-line marketing personnel. Due to the lack of a clear profile of potential customers, the marketing plan design is relatively rough and the form of reaching customers is single. In the marketing promotion stage, channel marketing personnel are not clear about the customer's purchase intention and need to repeatedly try different marketing logics to explore customer needs and achieve conversion.

[0003] However, the health insurance marketing methods in related technologies need to rely on manual experience to screen customers, which is not very accurate. When the customer range is limited, it will result in a waste of high-quality health insurance potential customer resources. Moreover, the marketing model is broad and not very targeted. Repeated outreach affects customer experience and lengthens the communication cycle, increasing marketing costs and losses while making it difficult to promote sales. It is difficult to maintain the team's enthusiasm for the sales of health insurance products, which needs to be solved urgently. Summary of the invention

[0004] The present application provides a method and device for identifying a user's intention to purchase an insurance product, so as to solve the problem that the health insurance marketing method in the related art needs to rely on manual experience to screen the customer base, which is not very accurate and will result in a waste of high-quality health insurance potential customer resources when the customer range is limited. In addition, the marketing model is broad and not very targeted. While repeated contact affects the customer experience, it also prolongs the communication cycle, increases marketing cost losses, makes it difficult to promote sales, and makes it difficult to maintain the team's enthusiasm for the sales of health insurance products.

[0005] The first aspect of the present application provides a method for identifying a user's intention to purchase an insurance product, which is applied to the model building stage and includes the following steps: collecting the user's identity identification and at least one of the corresponding historical insurance information, policy business information, complaint and consultation information, and customer portrait information; based on a preset customer vital sign system, generating a customer data set according to the user's identity identification and at least one of the corresponding historical insurance information, policy business information, complaint and consultation information, and customer portrait information; using the customer data set to build a customer classification model to determine the main characteristic factors that affect the customer's intention to purchase insurance; building a clustering model based on the main characteristic factors, and using the clustering model to output the target customer's actual insurance product purchase intention.

[0006] Optionally, in one embodiment of the present application, the main characteristic factors include at least one of the number of days until the most recent policy expiration, age, age of the first insured person, the first insurance type code and the number of customer policies.

[0007] Optionally, in an embodiment of the present application, it also includes: generating an insurance purchase intention customer group label and a corresponding customer portrait of the target customer based on the actual insurance product purchase intention.

[0008] Optionally, in an embodiment of the present application, it also includes: generating marketing activities and / or sales promotion words for the target customers based on the insurance purchasing intention customer group label and the corresponding customer portrait of the target customers.

[0009] Optionally, in one embodiment of the present application, before generating the customer data set, it also includes: based on at least one insurance company business scenario, generating a customer feature system including one or more dimensions of customer basic portrait, policy business, product features, customer type, value information, contact information, and risk information.

[0010] Optionally, in one embodiment of the present application, generating a customer data set based on the user's identity identifier and at least one of the corresponding historical insurance information, policy business information, complaint and consultation information, and customer portrait information includes: generating at least one user feature based on the user's identity identifier and at least one of the corresponding historical insurance information, policy business information, complaint and consultation information, and customer portrait information; performing missing value processing on the at least one user feature to obtain at least one feature that meets preset conditions; performing categorical feature conversion on a feature that is not a numerical type among the at least one feature to obtain the customer data set in combination with a feature that is a numerical type among the at least one feature.

[0011] The second aspect of the present application provides a device for identifying a user's intention to purchase an insurance product, which is applied to a model building stage and includes: a collection module for collecting a user's identity and at least one of the corresponding historical insurance information, policy business information, complaint and consultation information, and customer portrait information; a first generation module for generating a customer data set based on a preset customer vital sign system and according to the user's identity and at least one of the corresponding historical insurance information, policy business information, complaint and consultation information, and customer portrait information; a determination module for constructing a customer classification model using the customer data set to determine the main characteristic factors that affect the customer's intention to purchase insurance; and an output module for constructing a clustering model based on the main characteristic factors to output the target customer's actual insurance product purchase intention using the clustering model.

[0012] Optionally, in one embodiment of the present application, the main characteristic factors include at least one of the number of days until the most recent policy expiration, age, age of the first insured person, the first insurance type code and the number of customer policies.

[0013] Optionally, in one embodiment of the present application, it also includes: a second generation module, used to generate the insurance purchase intention customer group label and the corresponding customer portrait of the target customer based on the actual insurance product purchase intention.

[0014] Optionally, in one embodiment of the present application, it also includes: a third generation module, which is used to generate marketing activities and / or sales promotion words for the target customers based on the insurance purchasing intention customer group label and the corresponding customer portrait of the target customers.

[0015] Optionally, in one embodiment of the present application, it also includes: a fourth generation module, which is used to generate a customer feature system of one or more dimensions including customer basic portrait, policy business, product features, customer type, value information, contact information, and risk information based on at least one insurance company business scenario before generating the customer data set.

[0016] Optionally, in one embodiment of the present application, the first generation module includes: a generation unit, used to generate at least one user feature based on the user's identity and corresponding historical insurance information, policy business information, complaint and consultation information, and customer portrait information; a processing unit, used to perform missing value processing on the at least one user feature to obtain at least one feature that meets preset conditions; a conversion unit, used to perform categorical feature conversion on the features of the at least one feature that are not of numerical type, so as to obtain the customer data set in combination with the features of the at least one feature that are of numerical type.

[0017] A third aspect of the present application provides an electronic device, comprising: a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor executes the program to implement a method for identifying a user's intention to purchase an insurance product as described in the above embodiment.

[0018] A fourth aspect of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, which, when executed by a processor, implements the above method for identifying a user's intention to purchase an insurance product.

[0019] The fifth aspect of the present application provides a computer program product, including a computer program, which, when executed, is used to implement the above method for identifying a user's intention to purchase an insurance product.

[0020] The embodiment of the present application can collect a variety of information of users and combine it with a certain customer vital sign system to generate a customer data set, so as to use the customer data set to build a customer classification model and a clustering model to finally output the actual insurance product purchase intention of the target customer. Thus, it is realized to predict the purchase intention of insurance products in a scientific and quantitative way, which is convenient for marketing personnel to quickly screen out customers who are easier to convert, effectively shorten the marketing cycle, reduce marketing costs, and greatly improve the prediction accuracy. At the same time, the model can be deployed online, and the customer's intention to purchase health insurance can be predicted automatically and intelligently, reducing invalid communication and invalid contact, which greatly improves the insurance digital marketing ability and user experience. Thus, it solves the problem that the health insurance marketing method in the related technology needs to rely on manual experience to screen the customer base, the accuracy is not high, and when the customer range is limited, it will cause a waste of high-quality health insurance potential customer resources, and the marketing model is broad and not targeted. Repeated contact will affect the customer experience and will also lengthen the communication cycle, increase the marketing cost loss, but promote the order difficulty, and it is difficult to maintain the team's enthusiasm for the sales of health insurance products.

[0021] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0023] Figure 1 A flowchart of a method for identifying a user's intention to purchase an insurance product according to an embodiment of the present application;

[0024] Figure 2 This is a flowchart of a method for identifying a customer's intention to purchase health insurance based on machine learning in one embodiment of the present application;

[0025] Figure 3 A schematic diagram of the structure of a device for identifying a user's intention to purchase an insurance product according to an embodiment of the present application;

[0026] Figure 4 It is a schematic diagram of the structure of an electronic device provided according to an embodiment of the present application.

[0027] Reference numerals:

[0028] 10-Device for identifying user's intention to purchase insurance products: 100-collection module, 200-first generation module, 300-determination module and 400-output module; 401-memory, 402-processor and 403-communication interface. DETAILED DESCRIPTION

[0029] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0030] The following is a description of the method and device for identifying the user's intention to purchase insurance products in the embodiment of the present application with reference to the accompanying drawings. In view of the fact that the health insurance marketing method in the related art mentioned in the above background technology needs to rely on manual experience to screen the customer base, the accuracy is not high, and when the customer range is limited, it will cause a waste of high-quality health insurance potential customer resources, and the marketing model is broad and not targeted. Repeated contact affects the customer experience and prolongs the communication cycle, increases the marketing cost loss, but makes it difficult to promote orders, and is difficult to maintain the team's enthusiasm for the sales of health insurance products. The present application provides a method for identifying the user's intention to purchase insurance products. In this method, a variety of user information can be collected and combined with a certain customer vital sign system to generate a customer data set, so as to use the customer data set to build a customer classification model and a clustering model to finally output the target customer's actual insurance product purchase intention. As a result, the purchase intention of insurance products is predicted in a scientific and quantitative way, which facilitates marketing personnel to quickly screen out customers who are easier to convert, effectively shortens the marketing cycle, reduces marketing costs, and greatly improves the prediction accuracy. At the same time, the model can be deployed online to automatically and intelligently predict the customer's intention to purchase health insurance, reduce invalid communication and invalid contact, and greatly improve the insurance digital marketing capabilities and user experience. This solves the problem that the health insurance marketing method in the relevant technology needs to rely on manual experience to screen the customer base, which has low accuracy and will result in a waste of high-quality health insurance potential customer resources when the customer range is limited. In addition, the marketing model is broad and not very targeted. Repeated contact affects the customer experience and lengthens the communication cycle, increasing marketing costs and losses while making it difficult to promote sales and maintaining the team's enthusiasm for health insurance products.

[0031] Specifically, Figure 1 A flowchart of a method for identifying a user's intention to purchase an insurance product provided in an embodiment of the present application.

[0032] like Figure 1 As shown, the method for identifying the user's intention to purchase insurance products includes the following steps:

[0033] In step S101, the user's identity identification and at least one of the corresponding historical insurance information, policy business information, complaint and consultation information, and customer portrait information are collected.

[0034] It is understandable that the user's identity identifier here can be understood as an identifier that can identify the user, such as the user's name, ID card information, age, gender, etc. By collecting the user's identity identifier information, the user's identity can be verified to ensure the accuracy and security of the obtained user information, so as to collect the user's related insurance and other information based on the identity identifier and associate it with the customer's unique identification code.

[0035] In certain embodiments, when the present application collects the user's identity and the user's related insurance information, it includes but is not limited to at least one of the user's historical insurance information, policy business information, complaint and consultation information, and customer profile information. When collecting this information, it can be collected from, but is not limited to, enterprise-level systems such as core business systems, call center systems, product factories, and intelligent marketing service systems.

[0036] Among them, historical insurance information records the user's past insurance purchase behavior, including the insurance products purchased, the insured amount, the payment method, the protection period, the insurance channel and the insurance method, etc. This information is crucial for evaluating the user's insurance needs and purchasing power. By analyzing historical insurance information, we can understand the user's insurance preferences, risk tolerance and consumption habits, and thus more accurately predict their future purchasing intentions.

[0037] Policy business information here can be understood as insurance business touchpoint information such as insurance claims, such as policyholder changes, insured changes, insurance amount adjustments, etc. This information reflects the user's dynamic demand and management of insurance contracts. By analyzing policy business information, we can understand the user's continued attention to insurance products and demand changes, and then adjust marketing strategies and service plans.

[0038] Complaints and consultation information are feedback generated by users when using insurance products or receiving insurance services. This information is of great significance for evaluating the service quality and user satisfaction of insurance companies. By analyzing complaints and consultation information, we can find out the problems and confusions encountered by users in the process of purchasing and using insurance products, so as to improve products and services in a targeted manner and enhance user experience and satisfaction.

[0039] Customer portrait information is a description of the user's image based on multi-dimensional user data (such as basic information, transaction records, browsing behavior, etc.). This information is obtained through data mining and analysis technology, including the user's demographic attributes, purchasing preferences, activity, associations, etc. Customer portrait information provides insurance companies with more comprehensive and in-depth user insights, which helps with precision marketing and personalized services.

[0040] The embodiment of the present application can collect at least one of the user's identity identification and corresponding historical insurance information, policy business information, complaint and consultation information, customer portrait information and other multi-dimensional data, which helps insurance companies to more accurately identify people who intend to purchase health insurance, understand their needs and preferences, and provide more accurate products and services. At the same time, these data also provide strong support for the insurance company's risk management, marketing strategy formulation and customer service improvement.

[0041] Step S102, based on the preset customer vital signs system, generates a customer data set according to the user's identity identification and at least one of the corresponding historical insurance information, policy business information, complaint and consultation information, and customer portrait information.

[0042] It can be understood that the preset customer feature system here can be understood as a pre-set framework and method for describing and analyzing the multi-dimensional features of customers.

[0043] In other embodiments, after obtaining the user's identity and corresponding historical insurance information, policy business information, complaint and consultation information, customer portrait information and other information, in order to facilitate the identification and analysis of customer information, the present application may design a certain customer feature system and perform processing so as to generate a customer data set based on the processed customer features.

[0044] The embodiment of the present application can generate a certain customer data set. By analyzing the customer data set, it is helpful for enterprises to formulate personalized marketing strategies for different customer groups, thereby improving customer engagement and conversion rate. In addition, the customer data set can be used to predict customer needs, timely discover the deficiencies of products and services, and make improvements and optimizations to meet customer expectations and needs and improve customer satisfaction.

[0045] Optionally, in one embodiment of the present application, before generating a customer data set, it also includes: based on at least one insurance company business scenario, generating a customer feature system including one or more dimensions of customer basic portrait, policy business, product features, customer type, value information, contact information, and risk information.

[0046] Based on the relevant descriptions of other embodiments, it can be understood that the present application can generate a customer data set based on a certain customer feature system and in combination with various information of the user.

[0047] When establishing a customer feature system, the embodiments of the present application are mainly but not limited to designing a customer feature system including but not limited to seven dimensions, such as customer basic portrait, policy business, product features, customer type, value information, contact information, and risk information, based on the insurance company's business scenarios. After completing the feature processing, a customer data mart can be formed.

[0048] Among them, the customer basic portrait can include but is not limited to the user's basic information such as name, gender, age, occupation, work unit, etc.

[0049] Insurance policy business may include, but is not limited to, the user's insurance policy status such as the number and type of policies purchased, the policy effective date, the expiration date, the premium payment status, and the claim records such as the number of claims, the claim amount, the claim reason, the claim processing time, etc.

[0050] Product characteristics may include, but are not limited to, the customer's preference type for insurance products, the degree of attention paid to terms such as the insurance period, coverage, and compensation ratio, as well as their expectations and needs for insurance products.

[0051] Customer types can include but are not limited to the customer's actual situation such as income level, occupation and other factors to classify customers.

[0052] Value information may include, but is not limited to, whether the customer has purchased other insurance products, the customer's interest and demand for other insurance products, or the customer's loyalty to different types of products.

[0053] Contact information may include, but is not limited to, the communication channels between corporate sales personnel and users, such as telephone, email, text messages, customer preferences and satisfaction with different communication channels, and customer satisfaction with the contact.

[0054] Risk information may include, but is not limited to, the customer's risk level, such as the results of assessing the customer's credit risk level based on factors such as the customer's claims record and premium payment status, as well as the formulation of corresponding risk prevention and control measures based on the customer's potential risks.

[0055] Next, the feature processing process in the embodiment of the present application is further explained.

[0056] Optionally, in one embodiment of the present application, a customer data set is generated based on the user's identity identification and at least one of the corresponding historical insurance information, policy business information, complaint and consultation information, and customer portrait information, including: generating at least one user feature based on the user's identity identification and at least one of the corresponding historical insurance information, policy business information, complaint and consultation information, and customer portrait information; performing missing value processing on at least one user feature to obtain at least one feature that meets preset conditions; performing categorical feature conversion on a feature that is not a numerical type in at least one feature to obtain a customer data set in combination with a feature that is a numerical type in at least one feature.

[0057] It is understandable that the preset features here can be understood as certain conditions that the user feature values ​​used when generating the customer data set need to meet, for example, the numerical type of the feature is correct, the numerical range of the feature meets the requirements, etc. It can be set or adjusted by professionals in this field according to actual conditions or actual needs, and the embodiments of this application are only for illustrative purposes and are not specifically limited.

[0058] In the actual implementation process, before performing feature processing, the present application needs to generate at least one user feature based on the various information collected from the user, and then perform certain feature processing on the user feature. Among them, the feature processing of the present application embodiment for the customer mainly includes but is not limited to the two aspects of missing value processing and categorical feature conversion.

[0059] The missing value processing in the embodiment of the present application is mainly because the processed features may have missing values. For example, when the customer has no expired insurance policies in history, the feature "the number of days from the last insurance policy expiration" is a null value. At this time, it is necessary to process the missing values ​​with some logic to obtain features that meet certain conditions before entering the algorithm model for training. It should be noted that in the embodiment of the present application, the missing values ​​are not filled based on the business logic and the actual meaning of the features alone, but can be used but not limited to the XGBoost algorithm (eXtreme Gradient Boosting, a machine learning algorithm based on gradient boosting trees, which can combine multiple weak learners to improve the overall prediction performance) to carry out relevant processing with its own processing logic for missing values.

[0060] Furthermore, the embodiments of the present application can also perform certain processing on some features that are not of numerical type, cannot be directly input into the model, and need to be encoded before use, such as features such as "customer's most recent payment type".

[0061] In the embodiments of the present application, Count encoding (also known as count encoding, which is a frequency-based encoding method that can replace the value of each category feature with the number of times the value appears in the data set) and LeaveOneOut encoding (also known as leave-one-out encoding, which is a target variable-based encoding method that can replace the value of each category feature with the average value of the target variable when the value is excluded) are mainly but not limited to using two encoding methods to encode category features with more and fewer value types, respectively, according to the characteristics of the features and data verification results.

[0062] For example, for categorical features with a large number of possible values, such as "the city to which the last underwriting institution belongs" which can take one of hundreds of city names, the embodiments of the present application can be, but are not limited to, encoded using the Count method. The remaining categorical features with a small number of possible values ​​can be, but are not limited to, encoded using LeaveOneOut. Among them, the specific features encoded using LeaveOneOut in the embodiments of the present application can be expressed as follows: the first insurance main insurance type, the most recent insurance main insurance type, the last insurance channel, etc.

[0063] Step S103, constructing a customer classification model using the customer data set to determine the main characteristic factors that affect the customer's intention to purchase insurance, wherein the main characteristic factors include at least one of the number of days until the last policy expiration, age, age of the first insured person, first insurance type code and the number of customer policies.

[0064] In certain embodiments, after a customer data set is generated based on the collected user information and a certain customer feature system, the embodiments of the present application can use the customer data set to build a customer classification model, thereby determining the main characteristic factors that may affect the customer's intention to purchase insurance, which helps to understand how to formulate corresponding strategies for different users.

[0065] Among them, the customer classification model in the embodiment of the present application can be understood as a model for classifying customers according to their intention to purchase insurance products. And the main characteristic factors in the embodiment of the present application include but are not limited to at least one of the number of days since the last policy expired, age, age of the first insured person, first insurance type code and number of customer policies.

[0066] For example, the present application can construct a customer classification model to predict the customer's intention to purchase insurance products. When constructing the classification model, the present application embodiment can be, but is not limited to, taking any individual insurance policy under the name of an insured person, and the insured person's health insurance policy (including group insurance, supplementary insurance) and other policies within one year after the effective date as a positive sample, and taking any individual insurance policy under the name of an insured person, and the insured person's uninsured health insurance policy (including group insurance, supplementary insurance) and other policies within one year after the effective date as a negative sample, so as to combine positive samples and negative samples to construct a customer classification model.

[0067] Furthermore, the embodiments of the present application can also, but are not limited to, select the AUC indicator (Area Under the Curve, the area under the ROC curve, which is an important indicator for measuring the performance of the classification model) and the TopN hit rate (which can be understood as the probability of accurate prediction of the top N customers) as evaluation indicators to evaluate the constructed model, and finally select the XGBoost algorithm to construct the classification model.

[0068] The XGBoost in the embodiment of the present application can be, but is not limited to, improved on the basis of GBDT (Gradient Boosting DecisionTree, i.e., gradient boosting decision tree, which is an integrated learning method that uses decision tree as the basic model and gradually optimizes the performance of the model through the strategy of gradient boosting), which can effectively improve the performance of XGBoost and thus be applicable to a wider range. The core idea of ​​the improvement in the embodiment of the present application can be understood as constantly adding trees and constantly performing feature splitting to grow a tree. Each time a tree is added, it is actually learning a new function f(x) to fit the residual of the last prediction. When the training is completed and k trees are obtained, the score of a sample must be predicted, which can be understood as the feature of the sample will fall into a corresponding leaf node in each tree, and each leaf node corresponds to a score. Finally, it is only necessary to add up the scores corresponding to each tree to obtain the predicted value of the sample.

[0069] Furthermore, XGBoost in the embodiment of the present application also adds a regularization term to the cost function, and introduces shrinkage and column sampling to control the complexity of the model, reduce calculations, and prevent model overfitting.

[0070] The embodiment of the present application can construct a customer classification model through XGBoost, thereby realizing automatic learning of the processing strategy of missing values. The model has strong generalization ability and can output ranking according to the importance of the features of the classification model to obtain the main feature factors that affect the customer's intention to purchase insurance, thereby ensuring the practicality of the present application while greatly improving the efficiency of the present application.

[0071] Step S104, constructing a clustering model based on the main characteristic factors, so as to output the actual insurance product purchasing intention of the target customers using the clustering model.

[0072] As a possible implementation method, after constructing a customer classification model to obtain the customer's main characteristic factors, the embodiment of the present application can construct a clustering model based on the customer's main characteristic factors to use the clustering model to output the target customer's actual insurance product purchase intention.

[0073] Among them, target customers here can be understood as potential customer groups with insurance needs or purchasing behaviors determined based on the characteristics of products or services, market demand or own marketing strategies in insurance marketing. Actual insurance product purchase intention here can be understood as the subjective willingness and tendency of target customers when considering purchasing insurance products.

[0074] In the embodiment of the present application, when constructing a clustering model based on the main characteristic factors of the customer's intention to purchase insurance, it can be implemented by, but not limited to, selecting a K-means algorithm, etc. Specifically, different methods can be selected by professionals in this field according to actual conditions or actual needs. The embodiment of the present application is only for illustrative purposes and is not specifically limited.

[0075] The embodiments of the present application can use the model to output the actual insurance product purchase intention of the target customers, breaking through the limitations of previous customer screening based on operational experience, and using a scientific and quantitative method to predict the insurance product purchase intention score, thereby improving the accuracy of the prediction results. At the same time, it helps marketers to screen out customers who are easier to convert and improve the work experience of marketers.

[0076] Optionally, in an embodiment of the present application, it also includes: generating insurance purchase intention customer group labels and corresponding customer portraits of target customers based on actual insurance product purchase intentions.

[0077] In certain embodiments, after obtaining the target customer's actual insurance product purchasing intention, the present application can also utilize classification models and clustering models to combine the actual insurance purchasing intentions of different customers to generate insurance purchasing intention customer group labels and customer portraits corresponding to the target customer.

[0078] For example, the present application can use the output intention score of the classification model and the historical purchase conversion rate as the basis for grouping, and use the clustering model to divide the target customers into three categories of customers with high, medium and low purchase intentions according to the thresholds [0, 0.08], [0.08, 0.30], and [0.30, 1] as the classification threshold standards. Then, the output customer groups can be deployed in the form of labels on the intelligent marketing service system or other platforms that can operate customers. The front-end business can combine the insurance purchase intention customer group labels and the corresponding customer portraits for customer screening and marketing.

[0079] For example, for a group of target customers who have a high intention to purchase family protection products, the customer group label can be a focus on providing comprehensive protection for family members, including life insurance, health insurance, accident insurance, etc.; customer portraits can be but are not limited to the general age level, professional background, and purchase motivation of the customer group. For example, the age level is usually in the middle-aged and young stage, with elderly parents and young children, and they are the economic pillars of the family. Their professional backgrounds are mostly middle-level managers in enterprises or freelancers, with stable income but certain pressure; purchase motivations are mostly to ensure that family members can get financial support when misfortune occurs and to reduce the family burden.

[0080] The embodiment of the present application can generate insurance purchase intention customer group labels and corresponding customer portraits for target customers based on their actual insurance product purchase intentions, and the classification model and clustering model in the embodiment of the present application can be deployed online, while supporting both online and offline batch prediction tasks, and outputting "health insurance purchase intention" labels to the front-end business system, making it convenient for marketing personnel to quickly screen out customers who are easier to convert, effectively shortening the marketing cycle, and reducing marketing costs.

[0081] Optionally, in an embodiment of the present application, it also includes: generating marketing activities and / or sales promotion words for target customers based on the target customers' insurance purchasing intention customer group labels and corresponding customer portraits.

[0082] In some embodiments, the present application may also combine the target customers' insurance purchasing intention customer group labels and corresponding customer portraits to design targeted marketing activities and sales promotion techniques.

[0083] A customer group is labeled as middle-aged career elites, family pillars, with strong health awareness and future planning needs;

[0084] The customer profile is as follows: middle-aged professionals aged 35-50, in the rising or stable career stage, with a medium to high income level, able to afford certain insurance costs, married with children, the breadwinner of the family, concerned about their own and their family's health, and willing to invest in health:

[0085] Marketing activities:

[0086] Activity theme: "Protect the future, insure love" - ​​Exclusive insurance plan for middle-aged workplace elites

[0087] Activity content: (1) Online lectures: Invite financial experts and health consultants to hold online lectures on insurance planning, health management, and retirement planning; (2) One-on-one consultation: Provide one-on-one insurance consulting services and customize insurance plans based on the specific circumstances of the customer; (3) Limited-time discounts: Provide limited-time discounts or free additional protection for specific insurance products; (4) Family packages: Launch insurance packages that include family members, emphasizing the importance of overall family protection.

[0088] Promotional words:

[0089] Opening remarks: "Hello, Mr. Zhang, I am Xiao Li, an insurance consultant at XX Insurance Company. I understand that you are a middle-aged professional elite with a successful career and a happy family. Thank you very much for your valuable time. I called you today to introduce you to our 'Protect the Future, Insure Love' insurance plan designed for customers like you."

[0090] Demand exploration: "Mr. Zhang, as the breadwinner of the family, you must be very concerned about the health and future of yourself and your family. Our insurance plan not only covers comprehensive health protection, but also provides strong support for your retirement planning and children's education fund. Have you ever thought about how your family will face the future life in case of any unexpected events?"

[0091] Product highlights: "Our insurance products have the following highlights: First, they cover a wide range of conditions, including major diseases, accidental injuries, hospitalization and other protections; second, they are flexibly customized, and the insured amount and coverage can be adjusted according to your specific needs; third, they are cost-effective, with limited-time discounts and additional protections, so that every penny you spend is spent wisely. In addition, we have also launched a family package, so that the whole family can benefit from one insurance."

[0092] Sales promotion: "Mr. Zhang, considering your current family situation and future plans, I think this insurance plan is very suitable for you. If you purchase insurance now, you can not only enjoy a limited-time discount, but also provide long-term and stable protection for your family. What do you think? If you have any questions or need further information, I will be happy to provide you with detailed answers and customized solutions."

[0093] The embodiments of the present application can combine the target customers' insurance purchasing intention customer group labels and corresponding customer portraits to design targeted marketing activities and sales promotion techniques, optimize the content of marketing activities, reduce invalid contacts, and thus improve the C-end user experience.

[0094] The present application is described in detail below with reference to a specific embodiment.

[0095] Figure 2 This is a flowchart of a method for identifying a customer's intention to purchase health insurance based on machine learning according to an embodiment of the present application. Figure 2 As shown:

[0096] Step S201, collect user historical information data from various systems of the enterprise, process and integrate the data, and complete data collection.

[0097] Step S202: construct user features based on the collected user information, and perform data preprocessing to obtain enterprise-level information data to be trained.

[0098] Step S203: Train and optimize multiple types of algorithms respectively, and select the best model based on model performance. The health insurance purchase total customer identification model is trained based on the best performing XGBoost algorithm.

[0099] Step S204, using the Kmeans algorithm to cluster the actual insurance product purchase intention scores of the users output by the classification model to output customer group labels and customer portraits that can distinguish the purchase intentions of health insurance products.

[0100] Step S205, connect the customer grouping labels and customer portraits to the enterprise front-end business system, and the business can match customer groups through teacher selection labels to achieve precision marketing.

[0101] According to the method for identifying the user's intention to purchase insurance products proposed in the embodiment of the present application, a variety of information of the user can be collected and combined with a certain customer vital sign system to generate a customer data set, so as to use the customer data set to build a customer classification model and a clustering model to finally output the actual insurance product purchase intention of the target customer. Thus, it is realized to predict the purchase intention of insurance products in a scientific and quantitative way, which is convenient for marketing personnel to quickly screen out customers who are easier to convert, effectively shorten the marketing cycle, reduce marketing costs, and greatly improve the prediction accuracy. At the same time, the model can be deployed online, and the customer's intention to purchase health insurance can be predicted automatically and intelligently, reducing ineffective communication and ineffective contact, which greatly improves the insurance digital marketing capabilities and user experience. Thus, it solves the problem that the health insurance marketing method in the related technology needs to rely on manual experience to screen customer groups, the accuracy is not high, and when the customer range is limited, it will cause a waste of high-quality health insurance potential customer resources, and the marketing model is broad and not targeted. Repeated contact will affect the customer experience and will also lengthen the communication cycle, increase marketing cost losses, but promote orders. It is difficult to maintain the team's enthusiasm for the sale of health insurance products.

[0102] Next, a device for identifying a user's intention to purchase an insurance product according to an embodiment of the present application will be described with reference to the accompanying drawings.

[0103] Figure 3 It is a schematic diagram of the structure of a device for identifying a user's intention to purchase an insurance product according to an embodiment of the present application.

[0104] like Figure 3 As shown, the device 10 for identifying the user's purchase intention for insurance products includes: a collection module 100, a first generation module 200, a determination module 300 and an output module 400.

[0105] The collection module 100 is used to collect the user's identity and at least one of the corresponding historical insurance information, insurance policy business information, complaint and consultation information, and customer portrait information;

[0106] The first generation module 200 is used to generate a customer data set based on a preset customer vital sign system according to the user's identity identification and at least one of the corresponding historical insurance information, insurance policy business information, complaint and consultation information, and customer portrait information;

[0107] A determination module 300 is used to construct a customer classification model using the customer data set to determine the main characteristic factors that affect the customer's intention to purchase insurance;

[0108] The output module 400 is used to construct a clustering model based on the main characteristic factors, so as to output the actual insurance product purchasing intention of the target customers using the clustering model.

[0109] Optionally, in one embodiment of the present application, the main characteristic factors include at least one of the number of days until the most recent policy expiration, age, age of the first insured person, the first insurance type code and the number of customer policies.

[0110] Optionally, in one embodiment of the present application, it also includes: a second generation module.

[0111] Among them, the second generation module is used to generate the insurance purchase intention customer group label and corresponding customer portrait of the target customers based on the actual insurance product purchase intention.

[0112] Optionally, in one embodiment of the present application, it also includes: a third generation module.

[0113] Among them, the third generation module is used to generate marketing activities and / or sales promotion words for target customers based on the target customers' insurance purchasing intention customer group labels and corresponding customer portraits.

[0114] Optionally, in one embodiment of the present application, it also includes: a fourth generation module.

[0115] Among them, the fourth generation module is used to generate a customer feature system of one or more dimensions including customer basic portrait, policy business, product features, customer type, value information, contact information, and risk information based on at least one insurance company business scenario before generating a customer data set.

[0116] Optionally, in one embodiment of the present application, the first generation module includes: a generation unit, a processing unit and a conversion unit.

[0117] The generating unit is used to generate at least one user feature according to the user's identity identification and at least one of the corresponding historical insurance information, insurance policy business information, complaint and consultation information, and customer portrait information;

[0118] A processing unit, used to perform missing value processing on at least one user feature to obtain at least one feature that meets a preset condition;

[0119] The conversion unit is used to convert the feature of at least one feature that is not of a numerical type into a categorical feature, so as to obtain a customer data set by combining the feature of at least one feature that is of a numerical type.

[0120] It should be noted that the aforementioned explanation of the embodiment of the method for identifying a user's intention to purchase an insurance product is also applicable to the device for identifying a user's intention to purchase an insurance product in this embodiment, and will not be repeated here.

[0121] According to the device for identifying the user's intention to purchase insurance products proposed in the embodiment of the present application, a variety of information of the user can be collected and combined with a certain customer vital sign system to generate a customer data set, so as to use the customer data set to build a customer classification model and a clustering model to finally output the actual insurance product purchase intention of the target customer. Thus, it is realized to predict the purchase intention of insurance products in a scientific and quantitative way, which is convenient for marketing personnel to quickly screen out customers who are easier to convert, effectively shorten the marketing cycle, reduce marketing costs, and greatly improve the prediction accuracy. At the same time, the model can be deployed online, and the customer's intention to purchase health insurance can be predicted automatically and intelligently, reducing ineffective communication and ineffective contact, which greatly improves the insurance digital marketing capabilities and user experience. Thus, it solves the problem that the health insurance marketing method in the related technology needs to rely on manual experience to screen customer groups, the accuracy is not high, and when the customer range is limited, it will cause a waste of high-quality health insurance potential customer resources, and the marketing model is broad and not targeted. Repeated contact will affect the customer experience and will also lengthen the communication cycle, increase the marketing cost loss, but promote the order difficulty, and it is difficult to maintain the team's enthusiasm for the sales of health insurance products.

[0122] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device may include:

[0123] Memory 401 , processor 402 , and a computer program stored in the memory 401 and executable on the processor 402 .

[0124] When the processor 402 executes the program, the method for identifying the user's intention to purchase insurance products provided in the above embodiment is implemented.

[0125] Furthermore, the electronic device further comprises:

[0126] The communication interface 403 is used for communication between the memory 401 and the processor 402 .

[0127] The memory 401 is used to store computer programs that can be executed on the processor 402 .

[0128] The memory 401 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0129] If the memory 401, the processor 402 and the communication interface 403 are implemented independently, the communication interface 403, the memory 401 and the processor 402 can be connected to each other through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 4 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0130] Optionally, in a specific implementation, if the memory 401, the processor 402 and the communication interface 403 are integrated on a chip, the memory 401, the processor 402 and the communication interface 403 can communicate with each other through an internal interface.

[0131] The processor 402 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0132] An embodiment of the present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above method for identifying a user's intention to purchase an insurance product.

[0133] The embodiment of the present application also provides a computer program product, including a computer program, which can run computer instructions. When the computer instructions are executed by a processor, the method for identifying a user's intention to purchase an insurance product provided by the embodiment of the present application is implemented.

[0134] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0135] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0136] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.

[0137] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or N wirings (electronic devices), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways as necessary and then storing it in a computer memory.

[0138] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above embodiment, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented by hardware, as in another embodiment, it can be implemented by any one or a combination of multiple of the following technologies known in the art: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0139] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0140] In addition, each functional unit in each embodiment of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0141] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present application. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A method for identifying a user's intention to purchase an insurance product, characterized in that: Applied to the model building stage, wherein the method comprises the following steps: Collecting the user's identity and at least one of the corresponding historical insurance information, policy business information, complaint and consultation information, and customer profile information; Based on the preset customer vital signs system, a customer data set is generated according to the user's identity identifier and at least one of the corresponding historical insurance information, policy business information, complaint and consultation information, and customer portrait information; Using the customer data set to build a customer classification model to determine the main characteristic factors that affect the customer's intention to purchase insurance; A clustering model is constructed based on the main characteristic factors, so as to output the actual insurance product purchasing intention of the target customers using the clustering model.

2. The method for identifying a user's intention to purchase an insurance product according to claim 1, characterized in that: The main characteristic factors include at least one of the number of days until the most recent policy expires, age, age of the first insured person, first insured type code and the number of customer policies.

3. The method for identifying a user's intention to purchase an insurance product according to claim 1, characterized in that: Also includes: Generate the target customer's insurance purchase intention customer group label and corresponding customer portrait based on the actual insurance product purchase intention.

4. The method for identifying a user's intention to purchase an insurance product according to claim 1, characterized in that: Also includes: Marketing activities and / or sales promotion scripts for the target customers are generated based on the insurance purchasing intention customer group labels and corresponding customer portraits of the target customers.

5. The method for identifying a user's intention to purchase an insurance product according to claim 1, characterized in that: Before generating the customer data set, the method further includes: Based on at least one insurance company business scenario, a customer feature system is generated including one or more dimensions of customer basic portrait, policy business, product features, customer type, value information, contact information, and risk information.

6. The method for identifying a user's intention to purchase an insurance product according to claim 1, characterized in that: Generating a customer data set according to the user's identity identifier and at least one of corresponding historical insurance information, insurance policy business information, complaint and consultation information, and customer portrait information includes: Generate at least one user feature based on the user's identity identifier and at least one of the corresponding historical insurance information, policy business information, complaint and consultation information, and customer portrait information; Performing missing value processing on the at least one user feature to obtain at least one feature that meets a preset condition; The feature that is not of a numerical type among the at least one feature is converted into a categorical feature, so as to combine the feature that is of a numerical type among the at least one feature to obtain the customer data set.

7. A device for identifying a user's intention to purchase an insurance product, applied in the model building stage, characterized in that: include: A collection module, used to collect the user's identity and at least one of the corresponding historical insurance information, policy business information, complaint and consultation information, and customer portrait information; A generation module, for generating a customer data set based on a preset customer vital signs system and according to the user's identity identification and corresponding historical insurance information, insurance policy business information, complaint and consultation information, and customer portrait information; A determination module, for building a customer classification model using the customer data set to determine the main characteristic factors that affect the customer's intention to purchase insurance; The output module is used to construct a clustering model based on the main characteristic factors, so as to output the actual insurance product purchasing intention of the target customers using the clustering model.

8. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for identifying a user's intention to purchase an insurance product as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the method for identifying a user's intention to purchase an insurance product as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed, it is used to implement the method for identifying the user's intention to purchase an insurance product as described in any one of claims 1-6.

Images