Insurance product information pushing method and device, storage medium and terminal

Abstract

The application discloses an insurance product information pushing method and device, a storage medium and a terminal, relates to the technical field of data processing, and aims at solving the technical problem of how to improve user determination effectiveness and thus improve the pushing accuracy of insurance product information. Mainly includes: obtaining user transaction history data, user basic attribute data and context basic features of an insurance product; determining information cross features based on the user transaction history data, the user basic attribute data and the context basic features, and performing classification processing on the information cross features and user features of a user to be screened through a user classification model that has completed model training to determine target users after classification, wherein the user classification model is obtained by training based on an interaction time training sample set; and pushing product information of the insurance product to the target users.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method and apparatus for pushing insurance product information, a storage medium, and a terminal. Background Technology

[0002] With the rapid development of big data processing technology, data mining has become an indispensable data processing method in the Internet industry, especially for information push, where big data mining is used to screen users as targets for push notifications.

[0003] Currently, existing methods for screening users typically extract a single characteristic and then push specific insurance products to them. For example, selecting male users and pushing Ping An Fu insurance product information to all male users. However, this single-characteristic extraction method misses a large number of potential users, resulting in missed target audiences and significantly reducing the effectiveness of information push. Furthermore, pushing information solely based on specific insurance products fails to meet users' needs for flexible access to insurance product information, thus reducing the accuracy of insurance product information push. Summary of the Invention

[0004] One technical problem that this invention aims to solve is how to improve the validity of user identification, thereby improving the accuracy of insurance product information delivery.

[0005] According to one aspect of the present invention, a method for pushing insurance product information is provided, comprising:

[0006] Obtain user transaction history data, user basic attribute data, and contextual basic features of the insurance products;

[0007] Based on the user transaction history data, the user basic attribute data, and the contextual basic features, information cross features are determined, and the information cross features and user features of the users to be screened are classified using a user classification model that has been trained to determine the target users after classification. The user classification model is trained based on the interaction time training sample set.

[0008] The product information of the insurance product is pushed to the target user.

[0009] Furthermore, the step of determining the cross-features of information based on the user's transaction history data, the user's basic attribute data, and the contextual basic features includes:

[0010] Multiple basic feature objects to be cross-referenced are determined from the user transaction history data, the user basic attribute data, and the contextual basic features, respectively.

[0011] Perform feature crossing processing on the basic feature objects to be crossed respectively, and determine the crossing depth of the basic feature objects to be crossed after feature crossing processing;

[0012] If the cross depth matches a preset cross depth range, the information cross feature is obtained based on the feature cross processing.

[0013] Furthermore, before determining the cross depth corresponding to the basic feature object to be crossed after feature cross processing, the method further includes:

[0014] Determine the classification type of the target user, and calculate the feature importance parameters of the basic feature objects to be cross-referenced based on the classification type;

[0015] The basic feature objects to be crossed are filtered based on the aforementioned feature importance parameters to obtain the basic feature objects to be crossed for which the cross-cross depth is to be calculated.

[0016] Furthermore, determining multiple cross-feature objects from the user transaction history data, the user basic attribute data, and the contextual basic features includes:

[0017] Obtain product information of the insurance product, wherein the product information includes at least one of the following: product type, product transaction method, product push dimension information, and product release path information;

[0018] Based on the product information, feature similarity is calculated for the user's transaction history data, the user's basic attribute information, and the contextual basic features to obtain feature similarity values;

[0019] Based on the feature similarity value, the basic feature object to be cross-referenced is extracted from the user's transaction history data, the user's basic attribute data, and the contextual basic features.

[0020] Furthermore, before classifying the information cross-features and user features of the users to be screened using a user classification model that has already been trained, and determining the target users after classification, the method further includes:

[0021] Obtain the interaction paths of different users to be filtered, and determine the interaction time information corresponding to the interaction paths;

[0022] Retrieve the user samples and information cross samples corresponding to the interaction time information, and construct the interaction time training sample set based on the user samples and the information cross samples;

[0023] The initial classification model is trained based on the interaction time training sample set to obtain the user classification model.

[0024] Furthermore, the acquisition of user transaction history data, user basic attribute data, and contextual basic features of the insurance product includes:

[0025] Send transaction verification requests to each application client so that the application client can verify the insurance product transaction.

[0026] Once the application verifies the transaction of the insurance product, a data acquisition request is sent to the application so that the application responds to the data acquisition request and provides the user's transaction history data and user basic attribute data.

[0027] Furthermore, the step of pushing the insurance product information to the target user includes:

[0028] Based on a preset push response mapping relationship, the response parameters between the product information of the insurance product and the target user are determined. The preset push response mapping relationship is used to characterize the potential response relationship between different users and different product information.

[0029] Obtain a push strategy that matches the response parameters, and push the product information to the target user according to the push strategy.

[0030] According to one aspect of the present invention, an insurance product information push device is provided, comprising:

[0031] The acquisition module is used to acquire user transaction history data, user basic attribute data, and the contextual basic features of the insurance product.

[0032] The determination module is used to determine information cross features based on the user transaction history data, the user basic attribute data, and the contextual basic features, and to classify the information cross features and user features of the users to be screened using a user classification model that has been trained, thereby determining the target users after classification. The user classification model is trained based on the interaction time training sample set.

[0033] The push module is used to push product information of the insurance product to the target user.

[0034] Furthermore, the determining module is specifically used to determine multiple basic feature objects to be crossed from the user transaction history data, the user basic attribute data, and the context basic features; to perform feature crossing processing on the basic feature objects to be crossed respectively, and to determine the crossing depth corresponding to the basic feature objects to be crossed after feature crossing processing; if the crossing depth matches a preset crossing depth range, then the information crossing feature is obtained based on the feature crossing processing.

[0035] Furthermore, the device also includes:

[0036] The calculation module is used to determine the classification type of the target user and calculate the feature importance parameters of the basic feature object to be cross-referenced based on the classification type.

[0037] The filtering module is used to filter the basic feature objects to be crossed based on the feature importance parameter, so as to obtain the basic feature objects to be crossed for which the cross depth is to be calculated.

[0038] Furthermore, the determining module is also used to obtain product information of the insurance product, the product information including at least one of product type, product transaction method, product push dimension information, and product release path information; based on the product information, perform feature similarity calculation on the user transaction history data, the user basic attribute information, and the contextual basic features to obtain feature similarity values; and extract the basic feature objects to be cross-referenced from the user transaction history data, the user basic attribute data, and the contextual basic features based on the feature similarity values.

[0039] Furthermore, the device also includes: a construction module,

[0040] The acquisition module is also used to acquire the interaction paths of different users to be filtered, and to determine the interaction time information corresponding to the interaction paths.

[0041] The construction module is used to retrieve user samples and information cross samples corresponding to the interaction time information, and construct the interaction time training sample set based on the user samples and the information cross samples.

[0042] The training module is used to train the initial classification model based on the interaction time training sample set to obtain the user classification model.

[0043] Furthermore, the acquisition module is specifically used to send transaction verification requests to each application terminal so that the application terminal can verify the insurance product transaction; after the application terminal passes the transaction verification of the insurance product, a data acquisition request is sent to the application terminal so that the application terminal responds to the data acquisition request and feeds back the user transaction history data and user basic attribute data.

[0044] Furthermore, the push module is specifically used to determine the response parameters between the insurance product information and the target user based on a preset push response mapping relationship, wherein the preset push response mapping relationship is used to characterize the potential response relationship between different users and different product information; obtain a push strategy that matches the response parameters, and push the product information to the target user according to the push strategy.

[0045] According to another aspect of the present invention, a storage medium is provided, wherein at least one executable instruction is stored therein, the executable instruction causing a processor to perform an operation corresponding to the above-described insurance product information push method.

[0046] According to another aspect of the present invention, a terminal is provided, comprising: a processor, a memory, a communication interface, and a communication bus, wherein the processor, the memory, and the communication interface communicate with each other through the communication bus;

[0047] The memory is used to store at least one executable instruction, which causes the processor to perform the operation corresponding to the above-described insurance product information push method.

[0048] By employing the above-described technical solutions, the technical solutions provided by the embodiments of the present invention have at least the following advantages:

[0049] This invention provides a method and apparatus for pushing insurance product information, a computer storage medium, and a terminal. Compared with existing technologies, the embodiments of this invention acquire user transaction history data, user basic attribute data, and contextual basic features of the insurance product; determine information cross-features based on the user transaction history data, the user basic attribute data, and the contextual basic features; and classify the information cross-features and user features of the users to be screened using a user classification model that has been trained, thereby determining the target users after classification. The user classification model is trained based on an interaction time training sample set; and pushes the insurance product information to the target users. This achieves the goal of determining the purpose of pushing insurance product information by accurately classifying users, avoiding the loss of target users, ensuring accurate pushing of insurance product information, meeting users' needs for flexible access to insurance product information, and greatly improving the accuracy of insurance product information pushing.

[0050] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0051] The accompanying drawings, which form part of this specification, illustrate embodiments of the invention and, together with the description, serve to explain the principles of the invention.

[0052] The invention will be more clearly understood with reference to the accompanying drawings and the following detailed description, wherein:

[0053] Figure 1 This invention provides a flowchart of an insurance product information push method according to an embodiment of the present invention.

[0054] Figure 2 This diagram illustrates a block diagram of an insurance product information push device according to an embodiment of the present invention.

[0055] Figure 3 A schematic diagram of the structure of a terminal provided in an embodiment of the present invention is shown. Detailed Implementation

[0056] Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps set forth in these embodiments do not limit the scope of the invention.

[0057] At the same time, it should be understood that, for ease of description, the dimensions of the various parts shown in the accompanying drawings are not drawn according to actual scale.

[0058] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the invention or its application or use.

[0059] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and equipment should be considered part of the specification.

[0060] It should be noted that similar labels and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be discussed further in subsequent figures.

[0061] Embodiments of this invention can be applied to computer systems / servers that can operate with a wide range of other general-purpose or special-purpose computing system environments or configurations. Examples of well-known computing systems, environments, and / or configurations suitable for use with computer systems / servers include, but are not limited to: personal computer systems, server computer systems, thin clients, thick clients, handheld or laptop devices, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputer systems, mainframe computer systems, and distributed cloud computing environments that include any of the above systems, etc.

[0062] Computer systems / servers can be described in the general context of computer system executable instructions (such as program modules) executed by the computer system. Typically, program modules can include routines, programs, object programs, components, logic, data structures, etc., which perform specific tasks or implement specific abstract data types. Computer systems / servers can be implemented in distributed cloud computing environments, where tasks are performed by remote processing devices linked through a communication network. In distributed cloud computing environments, program modules can reside on local or remote computing system storage media, including storage devices.

[0063] Single feature extraction methods often miss a large number of potential users, leading to missed target audiences during information push notifications and significantly reducing the effectiveness of information push. Furthermore, pushing information solely based on specific insurance products fails to meet users' needs for flexible access to insurance product information, thus reducing the accuracy of insurance product information push notifications. This invention provides a method for pushing insurance product information, such as... Figure 1 As shown, the method includes:

[0064] 101. Obtain user transaction history data, user basic attribute data, and the contextual basic features of the insurance product.

[0065] In this embodiment of the invention, the current execution end, acting as the execution end for pushing insurance product information, can be a terminal device or a server-side device, such as a cloud server, etc., and this embodiment of the invention does not impose specific limitations. The current execution end first obtains the user's transaction history data, user basic attribute data, and the contextual basic features of the insurance product. At this time, the insurance product includes, but is not limited to, risk investment insurance products and disease claim insurance products customized and generated by insurance companies. The user transaction history data is used to represent all data content of the user's historical data in completing the insurance product transaction process, including but not limited to page browsing time, page entry depth (such as clicking the insurance button, exposure of the agreement pop-up window, entering the health declaration, and immediate insurance purchase), the number of times medical insurance, accident insurance, and critical illness insurance were purchased, and the number of times the insurance product was browsed, etc. The user's basic attribute data includes basic data such as user age, user gender, donation amount, whether they follow the insurance public account, whether they follow the insurance mini-program, whether they are a corporate credit user, city, city level, mobile phone model, number of donations, and number of times the mini-program was shared, etc. The contextual basic features are used to represent the features used in the transaction of the insurance product, including but not limited to mobile phone model, city, account information, etc., and this embodiment of the invention does not impose specific limitations.

[0066] 102. Based on the user's transaction history data, the user's basic attribute data, and the contextual basic features, determine the information cross features, and use the user classification model that has completed model training to classify the information cross features and the user features of the users to be screened, and determine the target users after classification.

[0067] In this embodiment of the invention, to improve the accuracy of user classification for targeted information push, the current execution end constructs cross-features based on user transaction history data, user basic attribute data, and contextual basic features. This cross-feature construction involves combining user transaction history data, user basic attribute data, and contextual basic features (i.e., multiplying two or more feature matrices). The depth of the cross-features is then used to determine the final cross-features. The user classification model is then used to classify the users to be screened and select the users to be pushed to. The user classification model is trained using an interaction time training sample set. This model can be trained on an initial classification model using methods such as convolutional neural networks or support vector machines to obtain a user classification model suitable for user classification. In this embodiment, the target users classified by the user classification model can include, but are not limited to, users with different purchasing power categories, users with different transaction categories, users of different genders, and users of different professions. This embodiment does not impose specific limitations.

[0068] It should be noted that the interaction time training sample set in this embodiment of the invention is constructed by filtering the interaction time information corresponding to the interaction paths in the transaction platforms of various insurance products, in order to improve the accuracy of user classification model classification. Furthermore, after classifying the users to be screened, users from one category can be selected as target users, or multiple categories can be selected as target users; this embodiment of the invention does not impose specific limitations.

[0069] 103. Push the product information of the insurance product to the target user.

[0070] In this embodiment of the invention, after the current execution terminal determines the target user to be pushed to, it pushes the product information of the insurance product to be pushed to the client used by the target user. At this time, if the target user is of one category, the product information can be pushed directly; if the target user is of multiple categories, product information matching the insurance product can be filtered and pushed. This embodiment of the invention does not make specific limitations. In addition, in order to improve the effectiveness of product information push, the current execution terminal can also push based on a push strategy. This embodiment of the invention does not make specific limitations. Furthermore, the pushed product information includes, but is not limited to, insurance product transaction information, product introduction information, and other insurance product information related to the insurance product. This embodiment of the invention does not make specific limitations.

[0071] In another embodiment of the invention, for further definition and explanation, the step of determining information cross-features based on the user transaction history data, the user basic attribute data, and the contextual basic features includes:

[0072] Multiple basic feature objects to be cross-referenced are determined from the user transaction history data, the user basic attribute data, and the contextual basic features, respectively.

[0073] Perform feature crossing processing on the basic feature objects to be crossed respectively, and determine the crossing depth of the basic feature objects to be crossed after feature crossing processing;

[0074] If the cross depth matches a preset cross depth range, the information cross feature is obtained based on the feature cross processing.

[0075] To improve the accuracy of information push by using feature cross-referencing to cater to different user strata, the current execution terminal first determines multiple basic feature objects to be cross-referencing from user transaction history data, user basic attribute data, and contextual basic features. These basic feature objects represent the feature basis for deep cross-referencing of users. At this point, user transaction history data, user basic attribute data, and a portion of the contextual basic features can be randomly selected as basic feature objects. Alternatively, user transaction history data, user basic attribute data, and contextual basic features can be matched and extracted based on a specific time period or specific text content. This embodiment of the invention does not impose specific limitations. After determining the basic feature objects to be cross-referencing, the current execution terminal performs feature cross-referencing processing on these objects, thus constructing multiple information cross-features. At this point, feature cross-processing is the process of combining multiple basic feature objects to obtain new features. In a specific application scenario, this can involve identifying multiple basic feature objects to be cross-processed from user transaction history data, user basic attribute data, and contextual basic features, and representing them in the form of entities and relationships in Featuretools. Here, an entity refers to a table or data frame in the dataset, and a relationship refers to the connection between entities, such as the association between data. Then, a cross-matrix is ​​generated based on the basic unit primitives (a feature generation function) of features in Featuretools. Specifically, primitives are divided into two categories: aggregation primitives and transformation primitives. Aggregation primitives are used to perform aggregation operations on the relationships between entities, including but not limited to calculations such as sum, mean, max, min, and count, such as calculating the number of insurance products purchased by a user, the average price, etc. This embodiment of the invention does not impose specific limitations. Transformation primitives are used to transform the attributes of entities, including but not limited to representations such as year, month, day, hour, and weekday. For example, information such as year, month, day, and hour can be extracted from a date and time attribute. This embodiment of the invention does not impose specific limitations. By combining different primitives, various complex features can be generated as the basic feature objects to be crossed after cross-feature processing.

[0076] It should be noted that, in order to improve the accuracy of user classification and thus the effectiveness of information push, the current execution end calculates the cross depth of the basic feature objects to be crossed after each feature cross-processing, and compares it with a preset cross-depth range. The preset cross-depth range represents the maximum number of features that can be crossed when generating cross features. For example, the preset cross-depth range can be set to 2-10, meaning that at most two features can be crossed; this embodiment of the invention does not impose a specific limitation. The preset cross-depth range can be configured according to the specific dataset and classification model, thereby reducing overfitting of the training data and improving the model's generalization ability.

[0077] In another embodiment of the invention, to further define and illustrate, before determining the cross depth corresponding to the basic feature object to be crossed after feature cross processing, the method further includes:

[0078] Determine the classification type of the target user, and calculate the feature importance parameters of the basic feature objects to be cross-referenced based on the classification type;

[0079] The basic feature objects to be crossed are filtered based on the feature importance parameter to obtain the basic feature objects to be crossed for which the crossing depth is to be calculated.

[0080] To improve the accuracy of user classification and avoid generating too many invalid features, the current execution end filters the objects to be cross-referenced. That is, before determining the cross-reference depth, it first determines the classification type of the target user to be classified, and then calculates the feature importance parameters of the objects to be cross-referenced based on this classification type. Specifically, the current execution end can automatically filter features based on the Gini Impurity criterion. In this case, the classification type represents the type of user that can be classified, and can be configured based on push requirements, including but not limited to purchasing power classification, gender classification, occupation classification, historical user type classification, etc., to calculate the feature importance parameters of the objects to be cross-referenced based on the classification type. This embodiment of the invention does not impose specific limitations. In a specific implementation scenario, the feature importance parameters can be calculated using the importance parameter calculation formula of the Gini Impurity criterion to measure the importance of a feature to the classification result. For example, the importance parameter calculation formula is as follows: Where k is the number of categories, and pi is the proportion of the number of categories of the i-th category to the total number of categories, thus obtaining the feature importance parameter.

[0081] It should be noted that after the current execution end calculates the feature importance parameters, it filters the basic feature objects to be cross-referenced. Specifically, it first calculates the frequency of each basic feature object in each classification type. In each classification node, it calculates the corresponding Gini Impurity gain value by traversing all basic feature objects to be cross-referenced. At this point, for each basic feature object to be cross-referenced, the range of the gain value can be divided into m discrete intervals. Based on the classification type of the samples in each discrete interval, the Gini Impurity importance parameter corresponding to the m split points of this basic feature object to be cross-referenced is calculated. Then, the average value of the Gini Impurity importance parameter among these m split points is taken as the gain value of this basic feature object to be cross-referenced. The basic feature object to be cross-referenced corresponding to the maximum gain value is selected as the split point, dividing the classification node in two. The feature filtering of the child nodes continues until the recursion termination condition is met, resulting in the filtered basic feature objects to be cross-referenced for calculating the cross-reference depth.

[0082] In another embodiment of the invention, for further definition and explanation, the step of determining multiple basic feature objects to be cross-referenced from the user transaction history data, the user basic attribute data, and the contextual basic features includes:

[0083] Obtain product information for the insurance product;

[0084] Based on the product information, feature similarity is calculated for the user's transaction history data, the user's basic attribute information, and the contextual basic features to obtain feature similarity values;

[0085] Based on the feature similarity value, the basic feature object to be cross-referenced is extracted from the user's transaction history data, the user's basic attribute data, and the contextual basic features.

[0086] To improve the accuracy of user classification and thus the effectiveness of insurance product information push to users, the current execution end first obtains the product information of the insurance product when determining the basic feature objects to be cross-referenced. This product information includes at least one of the following: product type, product transaction method, product push dimension information, and product release path information. The product type represents the type of insurance product, including but not limited to personal insurance and property insurance. The product transaction method includes but is not limited to digital rights transactions, electronic banking transactions, and cash transactions. The product push dimension information represents the channel dimension when pushing the product, including but not limited to social applications, insurance-specific applications, and SMS push notifications. The product release path information includes third-party platforms (such as social platforms and banking platforms) and insurance platforms, etc. This embodiment of the invention does not impose specific limitations. When obtaining product information, pre-configured product information for insurance products can be directly retrieved for direct use; this embodiment of the invention does not impose specific limitations.

[0087] It should be noted that after obtaining product information, the current execution end performs feature similarity calculations with user transaction history data, user basic attribute information, and contextual basic features to obtain feature similarity values ​​between each feature. Since product information includes at least one of product type, product transaction method, product push dimension information, and product release path information, when calculating similarity, the product type can be individually compared with user transaction history data, user basic attribute information, and contextual basic features for text similarity calculation. Then, the text similarity between user basic attribute information, contextual basic features, and at least one of product type, product transaction method, product push dimension information, and product release path information can be calculated separately to obtain the feature similarity value. Furthermore, the similarity calculation method in this embodiment can be achieved by converting word vectors and then calculating word vector similarity values ​​for each word-form feature; this embodiment does not impose specific limitations. After the feature similarity value is calculated, the current execution end extracts the basic feature objects to be crossed from the user's transaction history data, user basic attribute data, and contextual basic features based on the feature similarity value. That is, features with high similarity can be filtered out as basic feature objects to be crossed by setting a similarity threshold. This similarity threshold can be configured according to the feature selection requirements, and this embodiment of the invention does not make specific limitations.

[0088] In another embodiment of the invention, for further definition and explanation, before the step of classifying the information cross features and user features of the users to be screened using a user classification model that has completed model training, and determining the classified target users, the method further includes:

[0089] Obtain the interaction paths of different users to be filtered, and determine the interaction time information corresponding to the interaction paths;

[0090] Retrieve the user samples and information cross samples corresponding to the interaction time information, and construct the interaction time training sample set based on the user samples and the information cross samples;

[0091] The initial classification model is trained based on the interaction time training sample set to obtain the user classification model.

[0092] To improve the accuracy of insurance product information delivery by classifying users, the current execution terminal pre-builds an initial classification model and trains it. First, the current execution terminal obtains the interaction paths of different users to be screened. These interaction paths represent the paths users take to obtain insurance product information, such as social media platform paths, banking platform paths, etc. Simultaneously, it determines the interaction time information corresponding to each interaction path. The interaction time information represents the time a user spends browsing or interacting on the corresponding platform. For example, if a user logs in or chats on a social media platform for one hour, the interaction time information is one hour. This can be determined by sending an interaction time statistics request to the corresponding social media platform; this embodiment of the invention does not impose specific limitations. After determining the interaction time information, it retrieves the user samples corresponding to this interaction time information, as well as information cross-samples. The user samples are all users within the interaction time information, and the information cross-samples are the cross-grouped users determined after cross-validation based on all these users. At this time, cross-validation can be performed using a time-series cross-validation method; this embodiment of the invention does not impose specific limitations. Furthermore, an interaction time training sample set is constructed based on user samples and cross-referenced information samples. In a specific application scenario, this training sample set can include, but is not limited to, dividing the user's interaction time with the platform into five groups. The data from the first group is used as the training set, the data from the second group as the validation set, the data from the first two groups as the training set, and the data from the third group as the validation set, and so on, to complete the 5-fold cross-validation of the time series. Based on this interaction time training sample set, the initial classification model is trained to obtain a user classification model, significantly improving the accuracy and stability of the user classification model. Additionally, in this embodiment, the initial classification model can be a neural network model, a convolutional neural network model, a support vector machine model, etc., preferably an LGBM framework. This embodiment does not impose specific limitations.

[0093] In another embodiment of the invention, for further definition and explanation, the step of obtaining user transaction history data, user basic attribute data, and contextual basic features of the insurance product includes:

[0094] Send transaction verification requests to each application client so that the application client can verify the insurance product transaction.

[0095] Once the application verifies the transaction of the insurance product, a data acquisition request is sent to the application so that the application responds to the data acquisition request and provides the user's transaction history data and user basic attribute data.

[0096] To achieve feature cross-referencing based on user transaction history data, basic user attribute data, and contextual characteristics, in a specific application scenario, the current execution end, while acquiring the aforementioned data features, ensures data security by sending transaction verification requests to various application ends. These application ends include, but are not limited to, social media platforms and banking platforms that interact with the current execution end for insurance business and share user information. Upon receiving a transaction verification request, each application end verifies the user's transaction against the information pushed by the current execution end. Specifically, it determines whether the user is a defaulter or a user on a transaction blacklist within the application end, and whether the insurance product violates the application end's insurance business requirements. If the user is not a defaulter or a user on a transaction blacklist, and the insurance product does not violate the application end's insurance business requirements, the application end reports successful transaction verification. Upon successful insurance product transaction verification, the current execution end sends a data retrieval request to the application end, prompting the application end to respond with the user's transaction history data and basic user attribute data.

[0097] In another embodiment of the invention, for further definition and explanation, the step of pushing the insurance product information to the target user includes:

[0098] The response parameters between the insurance product information and the target user are determined based on a preset push response mapping relationship.

[0099] Obtain a push strategy that matches the response parameters, and push the product information to the target user according to the push strategy.

[0100] To improve the effectiveness of insurance product information push notifications, when the current execution terminal pushes information to categorized target users, it first determines the response parameters between the product information and the target user based on a preset push response mapping relationship. At this point, the response parameter characterizes the likelihood of the target user responding to the pushed insurance product. Furthermore, the preset push response mapping relationship characterizes the potential response relationships between different users and different product information; the larger the response relationship, the larger the response parameter. The response relationship is configured by insurance technicians, and the response parameter is generated proportionally within the 0-1 value range based on the response relationship, thus determining a response parameter for the target user. This embodiment of the invention does not impose specific limitations. Additionally, after determining the target user's response parameter, the current execution terminal obtains the push strategy for this response parameter. Different response parameters are pre-configured with different push strategies. For example, a larger response parameter indicates a higher likelihood of the user viewing and purchasing; therefore, the push strategy can correspond to a low-frequency push method to optimize the efficiency of product information push. In this embodiment of the invention, the push strategy characterizes the number of times, time, and method (such as SMS, platform advertisements, private messages, etc.) of sending product information to the target user; this embodiment of the invention does not impose specific limitations.

[0101] This invention provides a method for pushing insurance product information. Compared with existing technologies, this invention acquires user transaction history data, user basic attribute data, and contextual basic features of the insurance product. Based on the user transaction history data, user basic attribute data, and contextual basic features, it determines information cross-features and classifies the information cross-features and user features of the users to be screened using a user classification model that has been trained. The user classification model is trained based on an interaction time training sample set. The invention then pushes the insurance product information to the target users. This method achieves the goal of pushing insurance product information by accurately classifying users, avoiding the loss of target users, ensuring accurate pushing of insurance product information, meeting users' needs for flexible access to insurance product information, and greatly improving the accuracy of insurance product information push.

[0102] Furthermore, as a response to the above Figure 1 The implementation of the method shown in this invention provides an insurance product information push device, such as... Figure 2 As shown, the device includes:

[0103] The acquisition module 21 is used to acquire user transaction history data, user basic attribute data, and the contextual basic features of the insurance product.

[0104] The determination module 22 is used to determine information cross features based on the user transaction history data, the user basic attribute data, and the contextual basic features, and to classify the information cross features and user features of the users to be screened using a user classification model that has been trained, and to determine the target users after classification. The user classification model is trained based on the interaction time training sample set.

[0105] The push module 23 is used to push product information of the insurance product to the target user.

[0106] Furthermore, the determining module is specifically used to determine multiple basic feature objects to be crossed from the user transaction history data, the user basic attribute data, and the context basic features; to perform feature crossing processing on the basic feature objects to be crossed respectively, and to determine the crossing depth corresponding to the basic feature objects to be crossed after feature crossing processing; if the crossing depth matches a preset crossing depth range, then the information crossing feature is obtained based on the feature crossing processing.

[0107] Furthermore, the device also includes:

[0108] The calculation module is used to determine the classification type of the target user and calculate the feature importance parameters of the basic feature object to be cross-referenced based on the classification type.

[0109] The filtering module is used to filter the basic feature objects to be crossed based on the feature importance parameter, so as to obtain the basic feature objects to be crossed for which the cross depth is to be calculated.

[0110] Furthermore, the determining module is also used to obtain product information of the insurance product, the product information including at least one of product type, product transaction method, product push dimension information, and product release path information; based on the product information, perform feature similarity calculation on the user transaction history data, the user basic attribute information, and the contextual basic features to obtain feature similarity values; and extract the basic feature objects to be cross-referenced from the user transaction history data, the user basic attribute data, and the contextual basic features based on the feature similarity values.

[0111] Furthermore, the device also includes: a construction module,

[0112] The acquisition module is also used to acquire the interaction paths of different users to be filtered, and to determine the interaction time information corresponding to the interaction paths.

[0113] The construction module is used to retrieve user samples and information cross samples corresponding to the interaction time information, and construct the interaction time training sample set based on the user samples and the information cross samples.

[0114] The training module is used to train the initial classification model based on the interaction time training sample set to obtain the user classification model.

[0115] Furthermore, the acquisition module is specifically used to send transaction verification requests to each application terminal so that the application terminal can verify the insurance product transaction; after the application terminal passes the transaction verification of the insurance product, a data acquisition request is sent to the application terminal so that the application terminal responds to the data acquisition request and feeds back the user transaction history data and user basic attribute data.

[0116] Furthermore, the push module is specifically used to determine the response parameters between the insurance product information and the target user based on a preset push response mapping relationship, wherein the preset push response mapping relationship is used to characterize the potential response relationship between different users and different product information; obtain a push strategy that matches the response parameters, and push the product information to the target user according to the push strategy.

[0117] This invention provides an insurance product information push device. Compared with the prior art, this invention acquires user transaction history data, user basic attribute data, and contextual basic features of the insurance product. Based on the user transaction history data, user basic attribute data, and contextual basic features, it determines information cross features and classifies the information cross features and user features of the users to be screened using a user classification model that has been trained. The user classification model is trained based on an interaction time training sample set. The device pushes the insurance product information to the target users, achieving the goal of determining the push purpose of insurance product information through accurate user classification, avoiding the loss of target users, ensuring accurate push of insurance product information, meeting users' needs for flexible access to insurance product information, and greatly improving the accuracy of insurance product information push.

[0118] According to one embodiment of the present invention, a computer storage medium is provided, the storage medium storing at least one executable instruction, which can execute the insurance product information push method in any of the above method embodiments.

[0119] Figure 3 The diagram shows a structural schematic of a terminal according to an embodiment of the present invention. The specific embodiments of the present invention do not limit the specific implementation of the electronic device.

[0120] like Figure 3As shown, the terminal may include: a processor 302, a communications interface 304, a memory 306, and a communications bus 308.

[0121] The processor 302, communication interface 304, and memory 306 communicate with each other via communication bus 308.

[0122] Communication interface 304 is used to communicate with other network elements such as clients or other servers.

[0123] The processor 302 is used to execute program 310, specifically to execute the relevant steps in the above-described embodiment of the insurance product information push method.

[0124] Specifically, program 310 may include program code that includes computer operation instructions.

[0125] Processor 302 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention. The electronic device may include one or more processors of the same type, such as one or more CPUs; or it may include processors of different types, such as one or more CPUs and one or more ASICs.

[0126] Memory 306 is used to store program 310. Memory 306 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0127] Specifically, program 310 can be used to cause processor 302 to perform the following operations:

[0128] Obtain user transaction history data, user basic attribute data, and contextual basic features of the insurance products;

[0129] Based on the user transaction history data, the user basic attribute data, and the contextual basic features, information cross features are determined, and the information cross features and user features of the users to be screened are classified using a user classification model that has been trained to determine the target users after classification. The user classification model is trained based on the interaction time training sample set.

[0130] The product information of the insurance product is pushed to the target user.

[0131] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For system embodiments, since they largely correspond to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0132] The methods and systems of the present invention may be implemented in many ways. For example, they may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware. The above-described order of steps for the methods is for illustrative purposes only, and the steps of the methods of the present invention are not limited to the order specifically described above unless otherwise specifically stated. Furthermore, in some embodiments, the present invention may also be implemented as a program recorded on a recording medium, the program comprising machine-readable instructions for implementing the methods according to the present invention. Thus, the present invention also covers recording media storing programs for performing the methods according to the present invention.

[0133] The description of this invention is given for illustrative and descriptive purposes only and is not intended to be exhaustive or to limit the invention to the forms disclosed. Many modifications and variations will be apparent to those skilled in the art. The embodiments were chosen and described in order to better illustrate the principles and practical application of the invention and to enable those skilled in the art to understand the invention and to design various embodiments with various modifications suitable for a particular purpose.

Claims

1. An insurance product information push method, characterized by, include: Obtain user transaction history data, user basic attribute data, and contextual basic features of the insurance products; Based on the user transaction history data, the user basic attribute data, and the contextual basic features, information cross features are determined, and the information cross features and user features of the users to be screened are classified using a user classification model that has been trained to determine the target users after classification. The user classification model is trained based on the interaction time training sample set. Push the insurance product information to the target user; Before classifying the information cross features and user features of the users to be screened using a user classification model that has already been trained, and determining the target users after classification, the method further includes: Obtain the interaction paths of different users to be filtered, and determine the interaction time information corresponding to the interaction paths; Retrieve the user samples and information cross samples corresponding to the interaction time information, and construct the interaction time training sample set based on the user samples and the information cross samples; The initial classification model is trained based on the interaction time training sample set to obtain the user classification model.

2. The method of claim 1, wherein, The cross-features determined based on the user's transaction history data, the user's basic attribute data, and the contextual basic features include: Multiple basic feature objects to be cross-referenced are determined from the user transaction history data, the user basic attribute data, and the contextual basic features, respectively. Perform feature crossing processing on the basic feature objects to be crossed respectively, and determine the crossing depth of the basic feature objects to be crossed after feature crossing processing; If the cross depth matches a preset cross depth range, the information cross feature is obtained based on the feature cross processing.

3. The method of claim 2, wherein, Before determining the cross depth corresponding to the basic feature object to be crossed after feature cross processing, the method further includes: Determine the classification type of the target user, and calculate the feature importance parameters of the basic feature objects to be cross-referenced based on the classification type; The basic feature objects to be crossed are filtered based on the feature importance parameter to obtain the basic feature objects to be crossed for which the crossing depth is to be calculated.

4. The method of claim 2, wherein, The step of determining multiple cross-referenced basic feature objects from the user transaction history data, the user basic attribute data, and the contextual basic features includes: Obtain product information of the insurance product, wherein the product information includes at least one of the following: product type, product transaction method, product push dimension information, and product release path information; Based on the product information, feature similarity is calculated for the user's transaction history data, the user's basic attribute data, and the contextual basic features to obtain feature similarity values; Based on the feature similarity value, the basic feature object to be cross-referenced is extracted from the user's transaction history data, the user's basic attribute data, and the contextual basic features.

5. The method of claim 1, wherein, The acquisition of user transaction history data, user basic attribute data, and contextual basic features of the insurance product includes: Send transaction verification requests to each application client so that the application client can verify the insurance product transaction. Once the application verifies the transaction of the insurance product, a data acquisition request is sent to the application so that the application responds to the data acquisition request and provides the user's transaction history data and user basic attribute data.

6. The method according to any one of claims 1 to 5, characterized in that, The process of pushing the insurance product information to the target user includes: Based on a preset push response mapping relationship, the response parameters between the product information of the insurance product and the target user are determined. The preset push response mapping relationship is used to characterize the potential response relationship between different users and different product information. Obtain a push strategy that matches the response parameters, and push the product information to the target user according to the push strategy.

7. An insurance product information push device characterized by comprising: include: The acquisition module is used to acquire user transaction history data, user basic attribute data, and the contextual basic features of the insurance product. The determination module is used to determine information cross features based on the user transaction history data, the user basic attribute data, and the contextual basic features, and to classify the information cross features and user features of the users to be screened using a user classification model that has been trained, thereby determining the target users after classification. The user classification model is trained based on the interaction time training sample set. The push module is used to push product information of the insurance product to the target user; The device further includes: a construction module, The acquisition module is also used to acquire the interaction paths of different users to be filtered, and to determine the interaction time information corresponding to the interaction paths. The construction module is used to retrieve user samples and information cross samples corresponding to the interaction time information, and construct the interaction time training sample set based on the user samples and the information cross samples. The training module is used to train the initial classification model based on the interaction time training sample set to obtain the user classification model.

8. A computer storage medium storing at least one executable instruction that causes a processor to perform an operation corresponding to the insurance product information push method as described in any one of claims 1-6.

9. A terminal comprising: The processor, memory, communication interface, and communication bus are provided, wherein the processor, memory, and communication interface communicate with each other via the communication bus. The memory is used to store at least one executable instruction, which causes the processor to perform the operation corresponding to the insurance product information push method as described in any one of claims 1-6.

Images