Target insurance type purchase prediction method, device and equipment and storage medium thereof

Abstract

The embodiment of the invention belongs to the technical field of artificial intelligence, and relates to a target insurance type purchase prediction method, device and equipment and a storage medium thereof, and the method comprises the steps: obtaining first-class factor data which affects a target user to purchase a second insurance type from user portrait information of a first insurance type; obtaining purchase prediction factor data of the second insurance type; preprocessing the purchase prediction factor data of the second insurance type to obtain target prediction data; and inputting the target prediction data into a trained second insurance type purchase prediction model, and outputting second insurance type purchase data. The target insurance type purchase prediction method is applied to a cross-insurance type purchase prediction scene in financial insurance services, existing insurance type users are used as cross-insurance type service expansion groups, original customer groups are reasonably utilized, customer expansion processes are reduced to a certain extent, existing resources of companies are reasonably used, and moreover, the purchase prediction efficiency is improved. And business transverse development and expansion among different insurance types are facilitated.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology and is applied to the scenario of predicting cross-insurance type purchases in financial insurance business, and relates to a method, device, equipment and storage medium for predicting the purchase of a target insurance type. Background Art

[0002] With the rapid development of financial insurance business, in order to make the financial insurance business more and more adaptable to social progress, insurance institutions have launched more and more insurance type business.

[0003] Currently, for the business of different insurance types, they are processed separately. For example: for auto insurance business, there is a fixed business processing process in customer acquisition or subsequent business purchase; for health insurance business, there is also a different business processing process in customer acquisition or subsequent business, realizing vertical business specifications under different insurance types. However, this vertical business processing method is not conducive to the reasonable and scientific utilization of existing business data, and is also not conducive to the horizontal processing and expansion across insurance types. Summary of the Invention

[0004] The purpose of the embodiments of the present application is to propose a method, device, equipment and storage medium for predicting the purchase of a target insurance type, so as to solve the problem that the vertical business processing method in the existing financial insurance business field is not conducive to the reasonable and scientific utilization of existing business data, and is also not conducive to the horizontal processing and expansion across insurance types.

[0005] In a first aspect, the embodiments of the present application provide a method for predicting the purchase of a target insurance type, and adopt the following technical solutions:

[0006] A method for predicting the purchase of a target insurance type includes the following steps:

[0007] Obtain first type of factor data that affects a target user's purchase of a second insurance type from the user portrait information of a first insurance type, where the first insurance type represents a non-target insurance type, and the second insurance type represents a target insurance type;

[0008] Analyze the first type of factor data to determine whether the target user has the subject matter of the second insurance type;

[0009] If the target user has the subject matter of the second insurance type, obtain second type of factor data that affects the target user's purchase of the second insurance type;

[0010] Merge the first type of factor data and the second type of factor data to obtain second insurance type purchase prediction factor data;

[0011] If the target user does not have the subject matter of the second insurance type, only obtain the second insurance type purchase prediction factor data based on the first type of factor data;

[0012] Preprocess the second insurance type purchase prediction factor data to obtain target prediction data;

[0013] Input the target prediction data into the trained second insurance type purchase prediction model to output the second insurance type purchase data.

[0014] In a second aspect, an embodiment of the present application further provides a target insurance type purchase prediction device, which adopts the following technical solutions:

[0015] A target insurance type purchase prediction device includes:

[0016] A first type of factor data acquisition module, configured to obtain the first type of factor data that affects the target user's purchase of the second insurance type from the user portrait information of the first insurance type, where the first insurance type represents a non-target insurance type, and the second insurance type represents the target insurance type;

[0017] A first type of factor data analysis and judgment module, configured to analyze the first type of factor data to determine whether the target user has the subject matter of the second insurance type;

[0018] A first branch processing module, configured to, if the target user has the subject matter of the second insurance type, obtain the second type of factor data that affects the target user's purchase of the second insurance type;

[0019] A factor data merging module, configured to merge the first type of factor data and the second type of factor data to obtain the second insurance type purchase prediction factor data;

[0020] A second branch processing module, configured to, if the target user does not have the subject matter of the second insurance type, only obtain the second insurance type purchase prediction factor data according to the first type of factor data;

[0021] A target prediction data acquisition module, configured to preprocess the second insurance type purchase prediction factor data to obtain target prediction data;

[0022] A second insurance type purchase prediction module, configured to input the target prediction data into the trained second insurance type purchase prediction model to output the second insurance type purchase data.

[0023] In a third aspect, an embodiment of the present application further provides a computer device, which adopts the following technical solutions:

[0024] A computer device includes a memory and a processor. A computer-readable instruction is stored in the memory, and when the processor executes the computer-readable instruction, the steps of the above-mentioned target insurance type purchase prediction method are implemented.

[0025] Fourthly, an embodiment of the present application further provides a computer-readable storage medium, adopting the following technical solution:

[0026] A computer-readable storage medium has computer-readable instructions stored thereon, and when the computer-readable instructions are executed by a processor, the steps of the target insurance type purchase prediction method described above are implemented.

[0027] Compared with the prior art, the embodiments of the present application mainly have the following beneficial effects:

[0028] In the target insurance type purchase prediction method of the embodiment of the present application, by obtaining the first type of factor data that affects the target user's purchase of the second insurance type from the user portrait information of the first insurance type; obtaining the second insurance type purchase prediction factor data; preprocessing the second insurance type purchase prediction factor data to obtain the target prediction data; inputting the target prediction data into the trained second insurance type purchase prediction model, and outputting the second insurance type purchase data. Applying the target insurance type purchase prediction method to the cross-insurance type purchase prediction scenario in the financial insurance business, using the existing insurance type users as the business expansion group for cross-insurance types, compared with directly developing new users for the second insurance type from scratch, it reasonably utilizes the original customer group, reduces the customer acquisition process to a certain extent, reasonably uses the company's existing resources, and moreover, it is also convenient for horizontal business development and expansion between different insurance types. Description of the Drawings

[0029] In order to more clearly illustrate the solutions in the present application, the following will briefly introduce the drawings required for the description of the embodiments of the present application. Obviously, the following drawings are some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0030] Figure 1 is an exemplary system architecture diagram to which the present application can be applied;

[0031] Figure 2 is a flowchart of an embodiment of the target insurance type purchase prediction method according to the present application;

[0032] Figure 3 is a flowchart of a specific embodiment of initially supplementing the user portrait information of the first insurance type in the target insurance type purchase prediction method of the present application;

[0033] Figure 4 is Figure 3 a flowchart of a specific embodiment of step 302;

[0034] Figure 5It is a flowchart of a specific embodiment for secondary supplementation of user portrait information in the target insurance type purchase prediction method described in this application;

[0035] Figure 6 It is Figure 2 a flowchart of a specific embodiment of step 201;

[0036] Figure 7 It is Figure 2 a flowchart of a specific embodiment of step 206;

[0037] Figure 8 It is a flowchart of a specific embodiment for training the second insurance type purchase prediction model in the target insurance type purchase prediction method described in this application;

[0038] Figure 9 It is a schematic structural diagram of an embodiment of the target insurance type purchase prediction device according to this application;

[0039] Figure 10 It is a schematic structural diagram of an embodiment of the computer device according to this application. Detailed implementation manners

[0040] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which this application belongs; the terms used in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" and any variations thereof in the specification and claims of this application and the above drawings are intended to cover non-exclusive inclusion. The terms "first", "second", etc. in the specification and claims of this application or the above drawings are used to distinguish different objects and not to describe a specific order.

[0041] Reference to "embodiment" herein means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of this application. The phrase appears in various places in the specification and does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art will explicitly and implicitly understand that the embodiments described herein can be combined with other embodiments.

[0042] In order to enable those skilled in the technical field to better understand the solution of this application, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0043] Such as Figure 1As shown in the figure, the system architecture 100 may include a terminal device 101, a network 102, and a server 103. The terminal device 101 may be a laptop computer 1011, a tablet computer 1012, or a mobile phone 1013. The network 102 is a medium for providing a communication link between the terminal device 101 and the server 103. The network 102 may include various connection types, such as wired, wireless communication links, or fiber optic cables, etc.

[0044] Users can use the terminal device 101 to interact with the server 103 through the network 102 to receive or send messages, etc. Various communication client applications may be installed on the terminal device 101, such as a web browser application, a shopping application, a search application, an instant messaging tool, an email client, a social platform software, etc.

[0045] The terminal device 101 may be various electronic devices with a display screen and supporting web browsing. In addition to the laptop computer 1011, the tablet computer 1012, or the mobile phone 1013, the terminal device 101 may also be an e-book reader, an MP3 player (Moving Picture Experts Group Audio Layer III), an MP4 (Moving Picture Experts Group Audio Layer IV) player, a laptop portable computer, and a desktop computer, etc.

[0046] The server 103 may be a server that provides various services, such as a background server that provides support for the pages displayed on the terminal device 101.

[0047] It should be noted that the target insurance type purchase prediction method provided by the embodiments of the present application is generally executed by the server. Correspondingly, the target insurance type purchase prediction device is generally set in the server.

[0048] It should be understood that Figure 1 the numbers of the terminal devices, networks, and servers in

[0049] are merely illustrative. According to the implementation requirements, there may be any number of terminal devices, networks, and servers. Figure 2 Continuing to refer to

[0050] Step 201, obtain the first type of factor data that affects the target user's purchase of the second insurance type from the user portrait information of the first insurance type, where the first insurance type represents a non-target insurance type, and the second insurance type represents the target insurance type.

[0051] In this embodiment, the first insurance type and the second insurance type are different financial insurance products. For example, the first insurance type is vehicle insurance and the second insurance type is pet insurance; the first insurance type is vehicle insurance and the second insurance type is health insurance.

[0052] Specifically, taking the first insurance type as vehicle insurance and the second insurance type as pet insurance as an example, correspondingly, the user portrait information of the first insurance type, that is, the vehicle insurance user portrait information, and the first type of factor data refers to the characteristic factor data in the vehicle insurance user portrait information that affects the target user's purchase of pet insurance. The obtaining of the first type of factor data that affects the target user's purchase of the second insurance type is to obtain the characteristic factor data that affects the target user's purchase of pet insurance from the vehicle insurance user portrait information.

[0053] Generally, the vehicle insurance user portrait information includes the information of the owner's family members, such as single, having children, not having children, etc., and also includes the vehicle insurance user's historical insurance purchase behaviors, data reflecting the vehicle insurance user's insurance purchase awareness, etc. Since most single users like to raise pets, or when there are children at home, they choose not to raise pets as much as possible to avoid children coming into too much contact with allergens. Therefore, whether the vehicle owner is single or has children can be used as the first type of characteristic factor data for whether the vehicle insurance user purchases pet insurance. Moreover, the vehicle insurance user's historical insurance purchase behaviors and data reflecting the vehicle insurance user's insurance purchase awareness can be used as data reflecting the user's insurance purchase enthusiasm when purchasing insurance types. Therefore, such data can be used as auxiliary data for insurance purchase awareness to assist in analyzing the enthusiasm of vehicle insurance purchase users to purchase pet insurance.

[0054] Step 202: Analyze the first type of factor data to determine whether the target user has the subject matter of the second insurance type.

[0055] Specifically, continuing with the example where the first insurance type is vehicle insurance and the second insurance type is pet insurance for analysis, the subject matter of the second insurance type, that is, the subject matter of pet insurance, specifically refers to the pet itself. For example, a certain cat, a certain dog, etc.

[0056] Assume that the first type of factor data includes whether the vehicle insurance user raises a pet. Then, through this information, the judgment result has two situations. The first situation is that the vehicle insurance user raises a pet, and the second situation is that the vehicle insurance user does not raise a pet.

[0057] By analyzing the first type of factor data, it is determined whether the target user has the subject matter of the second insurance type, so as to facilitate the subsequent obtaining of the factor data affecting the purchase of the second insurance type.

[0058] Step 203: If the target user has the subject matter of the second insurance type, obtain the second type of factor data that affects the target user's purchase of the second insurance type.

[0059] Specifically, if the target user raises a certain pet, the second type of factor data refers to the data reflecting the subject matter, that is, the data of the pet itself, such as: pet breed, pet age, pet medical records, etc. The acquisition of the second type of factor data that affects the target user's purchase of the second insurance type is to obtain pet breed, pet age, and pet medical record information.

[0060] Step 204: Combine the first type of factor data and the second type of factor data to obtain the second insurance type purchase prediction factor data.

[0061] In this embodiment, the combination of the first type of factor data and the second type of factor data to obtain the second insurance type purchase prediction factor data means that in the case where the target user has the subject matter of the second insurance type, the first type of factor data and the second type of factor data are combined to obtain the comprehensive data for predicting the purchase of the second insurance type. In combination with the above examples of auto insurance and pet insurance, it means combining the first type of factor data obtained from auto insurance users and the second type of factor data obtained from the pet itself to obtain the comprehensive influence factor data for predicting the purchase of pet insurance.

[0062] Step 205: If the target user does not have the subject matter of the second insurance type, only obtain the second insurance type purchase prediction factor data based on the first type of factor data.

[0063] In this embodiment, the fact that the target user does not have the subject matter of the second insurance type can be understood as that the auto insurance user does not raise a pet. Obviously, the above-mentioned second type of factor data cannot be obtained. At this time, only obtain the second insurance type purchase prediction factor data based on the first type of factor data.

[0064] Step 206: Preprocess the second insurance type purchase prediction factor data to obtain the target prediction data;

[0065] In this embodiment, the preprocessing of the second insurance type purchase prediction factor data includes modifying obvious error data, supplementing, unifying the format, and vector representation, etc., to obtain the target prediction data with unified format and vector representation, avoiding inconsistent data formats input into the prediction model and reducing the prediction processing difficulty.

[0066] Step 207: Input the target prediction data into the trained second insurance type purchase prediction model and output the second insurance type purchase data.

[0067] In this embodiment, by obtaining the first type of factor data that affects the target user's purchase of the second insurance type from the user portrait information of the first insurance type; obtaining the second insurance type purchase prediction factor data; preprocessing the second insurance type purchase prediction factor data to obtain the target prediction data; inputting the target prediction data into the trained second insurance type purchase prediction model, and outputting the second insurance type purchase data. Applying the target insurance type purchase prediction method to the cross-insurance type purchase prediction scenario in the financial insurance business, using the existing insurance type users as the business expansion group for cross-insurance types, compared with directly developing the second insurance type users from scratch, it reasonably utilizes the original customer group, reduces the customer acquisition process to a certain extent, reasonably uses the company's existing resources, and moreover, it is also convenient for horizontal business development and expansion between different insurance types.

[0068] Continue to refer to Figure 3 , in some optional implementation manners, before executing step 201, the method further includes a step of initially supplementing the user portrait information of the first insurance type. Figure 3 FIG. is a flowchart of a specific embodiment of initially supplementing the user portrait information of the first insurance type in the target insurance type purchase prediction method described in this application, including the following steps:

[0069] Step 301, according to the page view monitoring program of the first insurance type business APP used by the target user, obtain the page view data and social sharing data of the target user;

[0070] Specifically, the target user refers to the user to be predicted for purchasing pet insurance, and the target user is an existing auto insurance user. The first insurance type business APP, for example, the owner service APP used by auto insurance users, and this owner service APP can be installed on the mobile phone of auto insurance users.

[0071] In this embodiment, according to the page view monitoring program of the first insurance type business APP used by the target user, obtaining the page view data and social sharing data of the target user can combine the page view data and social sharing data of auto insurance users to multi-dimensionally identify whether the user likes to browse pet-related information or the group data of pet breeders in the social circle, so as to be able to analyze the pet-raising tendency of this auto insurance user from the side and ensure the accuracy of subsequent pet insurance purchase analysis.

[0072] Step 302, based on the page view data and social sharing data, count the relevant browsing duration and sharing index of the target user for the subject matter corresponding to the second insurance type;

[0073] Step 303, use the relevant browsing duration and sharing index of the target user for the subject matter corresponding to the second insurance type as user portrait supplementary information, and supplement and cache it into the user portrait information of the target user.

[0074] Specifically, by using the relevant browsing duration and sharing index of the target user for the subject matter corresponding to the second insurance type as supplementary information for the user portrait, and caching it into the user portrait information of the target user, it is convenient to obtain a more comprehensive and multi-dimensional portrait information of the target user, ensuring that the subsequent obtained more comprehensive and multi-dimensional portrait information contains more data related to the prediction of the purchase of the second insurance type.

[0075] Continue to refer to Figure 4 , Figure 4 is Figure 3 The flowchart of a specific embodiment of step 302 includes:

[0076] Step 401, using natural language processing technology to analyze the page browsing data and the social sharing data, and analyzing the browsing behavior and social sharing behavior related to the subject matter of the second insurance type;

[0077] Specifically, using NLP natural speech recognition and processing technology, directly identify and extract pet-related information from the browsing page or social analysis page, so as to analyze the browsing behavior and social sharing behavior of the target user related to pets.

[0078] Step 402, based on the browsing behavior and the social sharing behavior, statistically calculate the relevant browsing duration and sharing index of the target user for the subject matter corresponding to the second insurance type.

[0079] By statistically calculating the relevant browsing duration and sharing index of the target user for the subject matter corresponding to the second insurance type, it is convenient to obtain a more comprehensive and multi-dimensional portrait information of the target user, ensuring that the subsequent obtained more comprehensive and multi-dimensional portrait information contains more data related to the prediction of the purchase of the second insurance type.

[0080] Continue to refer to Figure 5 , in some optional implementation manners, after performing step 303, the method further includes a step of secondary supplementation of the user portrait information. Figure 5 The flowchart of a specific embodiment of the secondary supplementation of the user portrait information in the target insurance type purchase prediction method described in this application includes the following steps:

[0081] Step 501, using a graph neural network to analyze the social relationship network of the target user. Among them, the graph neural network, such as a social relationship network based on GNN, contains the social relationship situation of the target user.

[0082] Step 502, according to the analysis result of the social relationship network, identify the number of first insurance type users with the subject matter of the second insurance type in the social relationship network of the target user, and the proportion value of the number of first insurance type users with the subject matter of the second insurance type in the social relationship network of the target user.

[0083] Specifically, by using a graph neural network to analyze the group of users who purchase vehicle insurance and raise pets in the target user's social relationship network, and to identify whether friends in the target user's social circle, etc., will affect their pet raising. Obviously, if there are more friends in the target user's social circle who raise pets and purchase pet insurance, the possibility of their purchase will be higher; if there are fewer friends in the target user's social circle who raise pets and purchase pet insurance, the possibility of their purchase will be lower, thereby improving the accuracy of subsequent pet insurance purchase prediction.

[0084] Step 503: Use the number of users of the first insurance type and the proportional value as supplementary information for the user portrait, and supplement and cache it into the user portrait information of the target user.

[0085] By using the number of users of the first insurance type and the proportional value as supplementary information for the user portrait, and supplementing and caching it into the user portrait information of the target user, it is convenient to obtain a more comprehensive and multi-dimensional portrait information of the target user, ensuring that the subsequent more comprehensive and multi-dimensional portrait information contains more data related to the prediction of the purchase of the second insurance type.

[0086] Continue to refer to Figure 6 , Figure 6 Yes Figure 2 It is a flowchart of a specific embodiment of step 201, including:

[0087] Step 601: Obtain the data tags that affect the purchase of the second insurance type set in advance;

[0088] Specifically, the data tags, for example: whether single, having children, not having children, browsing behaviors and social sharing behaviors related to pets, etc.

[0089] Step 602: Identify the cache library address information for storing the user portrait information of the first insurance type;

[0090] Step 603: Use the data tags and the identification information of the target user as the common retrieval fields, and the cache library address information as the target retrieval address, and screen out the cached data that affects the target user's purchase of the second insurance type from the local business data cache library as the first type of factor data.

[0091] In this embodiment, taking the existing vehicle insurance users as the business expansion group for cross-insurance types, rather than directly developing pet insurance users from scratch, reasonably utilizes the original customer group, reduces the customer acquisition process to a certain extent, and makes reasonable use of the company's existing resources.

[0092] In this embodiment, the step of obtaining the second type of factor data that affects the target user's purchase of the second insurance type specifically includes: obtaining, through questionnaire surveys or message feedback, the supplementary data input by the target user that affects the user's purchase of the second insurance type as the second type of factor data.

[0093] Specifically, when it is known that a car insurance user raises a pet, information such as the pet's breed, age, and health status of this car insurance user can be directly obtained through questionnaire surveys or message feedback.

[0094] Continue to refer to Figure 7 , Figure 7 Yes Figure 2 The flowchart of a specific embodiment of step 206 includes:

[0095] Step 701, perform deduplication processing, outlier processing, and data field format consistency processing on the second insurance type purchase prediction factor data to obtain the preprocessed second insurance type purchase influence data;

[0096] Specifically, it also includes missing value processing. The missing values in the second insurance type purchase prediction factor data can be filled using interpolation methods or K-nearest neighbor methods, or outlier replacement and correction processing can be performed using box plots, Z-scores, or interquartile ranges. The data field format consistency processing, for example: for different-dimensional data that may include time formats, numerical formats, string formats, etc., unified string format processing is performed to facilitate subsequent processing using a unified format transmission interface or processing interface.

[0097] Step 702, according to a preset vectorization representation dictionary, perform feature vectorization processing on each influence data in the preprocessed second insurance type purchase influence data to obtain the feature vectorization processing result;

[0098] Specifically, the step of performing feature vectorization processing on each influence data in the preprocessed second insurance type purchase influence data according to a preset vectorization representation dictionary refers to performing vectorization representation processing on the second insurance type purchase influence data that has undergone data field format consistency processing in combination with the vectorization representation dictionary, in order to facilitate subsequent direct use of the feature vectorization processing result for relevant predictions.

[0099] Step 703, set the feature vectorization processing result as the target prediction data.

[0100] Continue to refer to Figure 8 , in some optional implementation manners, before executing step 207, the method further includes a step of training the second insurance type purchase prediction model. Figure 8It is a flowchart of a specific embodiment for training the second insurance purchase prediction model in the target insurance purchase prediction method described in this application, including the following steps:

[0101] Step 801: Screen out the users who have purchased both the first insurance and the second insurance from the target business information cache library to generate a reference user group;

[0102] Specifically, the target business information cache library refers to the cache library used by financial insurance institutions to cache user portrait information of each insurance type, including self-built data warehouses, online data cloud platforms, etc.

[0103] Step 802: Obtain the user portrait information corresponding to each user in the reference user group, and the second insurance purchase data corresponding to each user;

[0104] Step 803: Analyze the user portrait information corresponding to each user in the reference user group to analyze the first type of factor data and the second type of factor data corresponding to each user;

[0105] Step 804: According to the user identification information, merge the corresponding first type of factor data and the second type of factor data, and obtain the second insurance purchase influence data corresponding to each user;

[0106] Step 805: According to the preset vectorization representation dictionary, perform feature vectorization processing on each influence data in the second insurance purchase influence data corresponding to each user to obtain the feature vectorization processing result;

[0107] Step 806: Based on the user identification information, perform grouped sorting on the second insurance purchase data and the feature vectorization processing result corresponding to different users to obtain the grouped sorting result as the model training data, where each piece of model training data includes the second insurance purchase data and the purchase influence data corresponding to a user;

[0108] Specifically, the second insurance purchase data and the purchase influence data corresponding to the same user are one piece of model training data. Assuming that there are 100 users in the reference user group, then corresponding to these 100 users, there are 100 pieces of model training data. The 100 pieces of model training data can be directly used as the model training set, or the 100 pieces of model training data can be divided into a training set, a test set, and a validation set to jointly learn and train the second insurance purchase prediction model to be trained.

[0109] By screening users who have purchased both the first insurance type and the second insurance type simultaneously, a reference user group is generated, and the first type of factor data and the second type of factor data corresponding to each user in the reference user group are analyzed to construct model training data. This ensures that the subsequent second insurance type purchase prediction model trained combines relatively comprehensive multi-dimensional influence data, making the prediction results more scientific and accurate during subsequent actual predictions.

[0110] Step 807: Input the model training data into the second insurance type purchase prediction model to be trained, learn the weight relationship between the second insurance type purchase data and the purchase influence data, and obtain the trained second insurance type purchase prediction model.

[0111] Specifically, the second insurance type purchase prediction model is constructed by combining a deep neural network (DNN) and a long short-term memory network (LSTM). During the model training process, the backpropagation algorithm can be used to adjust the network parameters to obtain the minimized prediction error. At the same time, an attention mechanism is introduced to enable the model to focus on the feature factor data in the user data that is most relevant to the pet insurance purchase decision, improving the prediction accuracy.

[0112] In this embodiment, the step of inputting the target prediction data into the trained second insurance type purchase prediction model and outputting the second insurance type purchase data specifically includes: based on the target prediction data, identifying the second insurance type purchase influence data corresponding to the target user; calculating according to the weight relationship between the second insurance type purchase data and the purchase influence data to obtain the second insurance type purchase data corresponding to the target user, where the calculation includes weighted accumulation summation calculation.

[0113] This application obtains the first type of factor data that affects the target user's purchase of the second insurance type from the user portrait information of the first insurance type; obtains the second insurance type purchase prediction factor data; preprocesses the second insurance type purchase prediction factor data to obtain the target prediction data; inputs the target prediction data into the trained second insurance type purchase prediction model and outputs the second insurance type purchase data. Applying the target insurance type purchase prediction method to the cross-insurance type purchase prediction scenario in the financial insurance business, using the existing insurance type users as the business expansion group for cross-insurance types, compared with directly developing the second insurance type users from scratch, reasonably utilizes the original customer group, reduces the customer acquisition process to a certain extent, rationally uses the company's existing resources, and is also convenient for horizontal business development and expansion between different insurance types.

[0114] Embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, Artificial Intelligence (AI) is a theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to obtain the best results.

[0115] Artificial intelligence basic technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, and mechatronics. Artificial intelligence software technologies mainly include several major directions such as computer vision technology, robotics, biometric technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0116] In the embodiments of the present application, by obtaining the first type of factor data that affects the target user's purchase of the second insurance type from the user portrait information of the first insurance type; obtaining the second insurance type purchase prediction factor data; preprocessing the second insurance type purchase prediction factor data to obtain the target prediction data; inputting the target prediction data into the trained second insurance type purchase prediction model, and outputting the second insurance type purchase data. Applying the target insurance type purchase prediction method to the cross-insurance type purchase prediction scenario in the financial insurance business, using the existing insurance type users as the cross-insurance type business expansion group, compared with directly developing the second insurance type users from scratch, reasonably utilizing the original customer group, reducing the customer acquisition process to a certain extent, reasonably using the company's existing resources, and moreover, facilitating horizontal business development and expansion between different insurance types.

[0117] Further referring to Figure 9 As an implementation of the method shown above Figure 2 In the present application, an embodiment of the target insurance type purchase prediction device is provided. This device embodiment corresponds to the method embodiment shown in Figure 2 and this device can be specifically applied to various electronic devices.

[0118] As Figure 9 shown, the target insurance type purchase prediction device 900 described in this embodiment includes: a first type of factor data acquisition module 901, a first type of factor data analysis and judgment module 902, a first branch processing module 903, a factor data merging module 904, a second branch processing module 905, a target prediction data obtaining module 906, and a second insurance type purchase prediction module 907. Among them:

[0119] The first type of factor data acquisition module 901 is used to obtain the first type of factor data that affects the target user's purchase of the second insurance type from the user portrait information of the first insurance type, where the first insurance type represents a non-target insurance type, and the second insurance type represents the target insurance type;

[0120] The first type of factor data analysis and judgment module 902 is used to analyze the first type of factor data and judge whether the target user has the subject matter of the second insurance type;

[0121] The first branch processing module 903 is used to obtain the second type of factor data that affects the target user's purchase of the second insurance type if the target user has the subject matter of the second insurance type;

[0122] The factor data merging module 904 is used to merge the first type of factor data and the second type of factor data to obtain the prediction factor data for the purchase of the second insurance type;

[0123] The second branch processing module 905 is used to only obtain the prediction factor data for the purchase of the second insurance type based on the first type of factor data if the target user does not have the subject matter of the second insurance type;

[0124] The target prediction data obtaining module 906 is used to preprocess the prediction factor data for the purchase of the second insurance type to obtain the target prediction data;

[0125] The second insurance type purchase prediction module 907 is used to input the target prediction data into the trained second insurance type purchase prediction model and output the second insurance type purchase data.

[0126] In this application, by obtaining the first type of factor data that affects the target user's purchase of the second insurance type from the user portrait information of the first insurance type; obtaining the prediction factor data for the purchase of the second insurance type; preprocessing the prediction factor data for the purchase of the second insurance type to obtain the target prediction data; inputting the target prediction data into the trained second insurance type purchase prediction model and outputting the second insurance type purchase data. Applying the target insurance type purchase prediction method to the cross-insurance type purchase prediction scenario in the financial insurance business, using the existing insurance type users as the business expansion group for cross-insurance types, compared with directly developing new users for the second insurance type from scratch, it reasonably utilizes the original customer group, reduces the customer acquisition process to a certain extent, reasonably uses the company's existing resources, and moreover, it is also convenient for horizontal business development and expansion between different insurance types.

[0127] In this embodiment, the target insurance type purchase prediction device 900 further includes: a user portrait supplementary information acquisition module, a user portrait supplementary information analysis module, and a user portrait supplementary information caching module. Among them:

[0128] The user portrait supplementary information acquisition module is used to obtain the page browsing data and social sharing data of the target user according to the page browsing monitoring program of the first insurance type business APP used by the target user; and is also used to analyze the social relationship network of the target user by using a graph neural network;

[0129] The user profile supplementary information analysis module is used to count the relevant browsing duration and sharing index of the target user for the subject matter corresponding to the second insurance type based on the page browsing data and social sharing data; according to the social relationship network analysis result, identify the number of first insurance type users with the subject matter of the second insurance type in the social relationship network of the target user, and the proportion value of the number of first insurance type users with the subject matter of the second insurance type in the social relationship network of the target user.

[0130] The user profile supplementary information caching module is used to supplement and cache the relevant browsing duration and sharing index of the target user for the subject matter corresponding to the second insurance type as user profile supplementary information into the user profile information of the target user; supplement and cache the number of first insurance type users and the proportion value as user profile supplementary information into the user profile information of the target user.

[0131] In this embodiment, the user profile supplementary information analysis module includes a natural language processing unit and a statistics unit. Among them:

[0132] The natural language processing unit is used to analyze the page browsing data and the social sharing data by using natural language processing technology, and analyze the browsing behavior and social sharing behavior related to the subject matter of the second insurance type.

[0133] The statistics unit is used to count the relevant browsing duration and sharing index of the target user for the subject matter corresponding to the second insurance type based on the browsing behavior and the social sharing behavior.

[0134] In this embodiment, the first type of factor data acquisition module 901 includes: a data label acquisition unit, an address information identification unit, and a retrieval and screening unit. Among them:

[0135] The data label acquisition unit is used to acquire the pre-set data labels that affect the purchase of the second insurance type.

[0136] The address information identification unit is used to identify the cache library address information for storing the user profile information of the first insurance type.

[0137] The retrieval and screening unit is used to use the data label and the identification information of the target user as the common retrieval fields, and the cache library address information as the target retrieval address, and screen out the cached data that affects the target user's purchase of the second insurance type from the local business data cache library as the first type of factor data.

[0138] In this embodiment, the target prediction data acquisition module 906 includes: a data preprocessing unit, a data vectorization representation unit, and a target prediction data setting unit. Among them:

[0139] A data preprocessing unit, configured to perform deduplication processing, outlier processing, and data field format consistency processing on the second insurance type purchase prediction factor data, to obtain the preprocessed second insurance type purchase impact data;

[0140] A data vectorization representation unit, configured to perform feature vectorization processing on each impact data in the preprocessed second insurance type purchase impact data according to a preset vectorization representation dictionary, to obtain a feature vectorization processing result;

[0141] A target prediction data setting unit, configured to set the feature vectorization processing result as the target prediction data.

[0142] In this embodiment, the target insurance type purchase prediction device 10 further includes: a reference user group screening module, a reference user group data acquisition module, a pre-training factor data analysis module, a pre-training factor data merging and processing module, a pre-training vectorization processing module, a model training data obtaining module, and a prediction model learning and training module. Among them:

[0143] The reference user group screening module is configured to screen out users who have purchased both the first insurance type and the second insurance type from the target business information cache library, and generate a reference user group;

[0144] The reference user group data acquisition module is configured to acquire the user portrait information corresponding to each user in the reference user group, and the second insurance type purchase data corresponding to each user;

[0145] The pre-training factor data analysis module is configured to analyze the user portrait information corresponding to each user in the reference user group, and analyze the first type of factor data and the second type of factor data corresponding to each user;

[0146] The pre-training factor data merging and processing module is configured to merge the corresponding first type of factor data and the second type of factor data according to the user identification information, and acquire the second insurance type purchase impact data corresponding to each user;

[0147] The pre-training vectorization processing module is configured to perform feature vectorization processing on each impact data in the second insurance type purchase impact data corresponding to each user according to a preset vectorization representation dictionary, to obtain a feature vectorization processing result;

[0148] The model training data obtaining module is configured to perform grouped sorting on the second insurance type purchase data and the feature vectorization processing result corresponding to different users based on the user identification information, to obtain a grouped sorting result as model training data, where each piece of model training data includes the second insurance type purchase data and the purchase impact data corresponding to a user;

[0149] A prediction model learning and training module for inputting the model training data into a second insurance type purchase prediction model to be trained, learning the weight relationship between the second insurance type purchase data and the purchase influencing data, and obtaining a trained second insurance type purchase prediction model.

[0150] In this embodiment, the second insurance type purchase prediction module 907 includes a second insurance type purchase influencing data identification unit and a second insurance type purchase data calculation unit. Specifically:

[0151] The second insurance type purchase influencing data identification unit is used to identify the second insurance type purchase influencing data corresponding to the target user based on the target prediction data;

[0152] The second insurance type purchase data calculation unit is used to calculate according to the weight relationship between the second insurance type purchase data and the purchase influencing data to obtain the second insurance type purchase data corresponding to the target user, where the calculation includes weighted accumulation summation calculation.

[0153] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a magnetic disk, an optical disk, a Read-Only Memory (ROM), or a Random Access Memory (RAM), etc.

[0154] It should be understood that although the steps in the flowchart of the accompanying drawings are shown in sequence according to the arrows, these steps do not necessarily have to be executed in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps has no strict order limit, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings can include multiple sub-steps or multiple stages. These sub-steps or stages do not necessarily have to be executed at the same time, but can be executed at different times, and their execution order does not necessarily have to be sequential, but can be executed alternately or alternately with at least a part of other steps or sub-steps or stages of other steps.

[0155] To solve the above technical problems, an embodiment of the present application also provides a computer device. For details, please refer to Figure 10 , Figure 10 which is the basic structural block diagram of the computer device in this embodiment.

[0156] The computer device 10 includes a memory 10a, a processor 10b, and a network interface 10c that are communicatively connected to each other via a system bus. It should be noted that Figure 10 only the computer device 10 with components such as a memory 10a, a processor 10b, and a network interface 10c is shown, but it should be understood that it is not required to implement all the shown components, and more or fewer components can be implemented alternatively. Among them, those skilled in the art of this technology can understand that a computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), embedded devices, etc.

[0157] The computer device can be a computing device such as a desktop computer, a notebook, a palm computer, and a cloud server. The computer device can perform human-computer interaction with a user through means such as a keyboard, a mouse, a remote control, a touchpad, or a voice control device.

[0158] The memory 10a includes at least one type of readable storage medium, and the readable storage medium includes flash memory, a hard disk, a multimedia card, a card-type memory (such as an SD or DX memory, etc.), a random access memory (RAM), a static random access memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, an optical disk, etc. In some embodiments, the memory 10a can be an internal storage unit of the computer device 10, such as the hard disk or memory of the computer device 10. In other embodiments, the memory 10a can also be an external storage device of the computer device 10, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the computer device 10. Of course, the memory 10a can also include both the internal storage unit and the external storage device of the computer device 10. In this embodiment, the memory 10a is generally used to store the operating system and various application software installed on the computer device 10, such as computer-readable instructions for a target insurance type purchase prediction method, etc. In addition, the memory 10a can also be used to temporarily store various data that have been output or will be output.

[0159] In some embodiments, the processor 10b may be a Central Processing Unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips. The processor 10b is generally used to control the overall operation of the computer device 10. In this embodiment, the processor 10b is used to run the computer-readable instructions stored in the memory 10a or process data, such as running the computer-readable instructions of the target insurance type purchase prediction method.

[0160] The network interface 10c may include a wireless network interface or a wired network interface. The network interface 10c is generally used to establish a communication connection between the computer device 10 and other electronic devices.

[0161] The computer device proposed in this embodiment belongs to the field of artificial intelligence technology and is applied to the cross-insurance type purchase prediction scenario in the financial insurance business. In this application, the first type of factor data that affects the target user's purchase of the second insurance type is obtained from the user portrait information of the first insurance type; the second insurance type purchase prediction factor data is obtained; the second insurance type purchase prediction factor data is preprocessed to obtain the target prediction data; the target prediction data is input into the trained second insurance type purchase prediction model, and the second insurance type purchase data is output. The target insurance type purchase prediction method is applied to the cross-insurance type purchase prediction scenario in the financial insurance business. Using the existing insurance type users as the business expansion group for cross-insurance types, compared with directly developing new users for the second insurance type from scratch, the original customer group is reasonably utilized, the customer acquisition process is reduced to a certain extent, the company's existing resources are reasonably used, and it is also convenient to carry out horizontal business development and expansion between different insurance types.

[0162] This application also provides another implementation manner, that is, a computer-readable storage medium is provided. The computer-readable storage medium stores computer-readable instructions, and the computer-readable instructions can be executed by a processor to enable the processor to execute the steps of the target insurance type purchase prediction method as described above.

[0163] The computer-readable storage medium proposed in this embodiment belongs to the field of artificial intelligence technology and is applied to the cross-insurance purchase prediction scenario in the financial insurance business. In this application, the first type of factor data that affects the target user's purchase of the second insurance is obtained from the user portrait information of the first insurance; the second insurance purchase prediction factor data is obtained; the second insurance purchase prediction factor data is preprocessed to obtain the target prediction data; the target prediction data is input into the trained second insurance purchase prediction model, and the second insurance purchase data is output. Applying the target insurance purchase prediction method to the cross-insurance purchase prediction scenario in the financial insurance business, using the existing insurance users as the business expansion group for cross-insurance, rather than directly developing the second insurance users from scratch, reasonably utilizes the original customer group, reduces the customer acquisition process to a certain extent, rationally uses the company's existing resources, and is also convenient for horizontal business development and expansion between different insurances.

[0164] Through the description of the above embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus a necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions for causing a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in various embodiments of this application.

[0165] Obviously, the above-described embodiments are only a part of the embodiments of this application, rather than all of the embodiments. The accompanying drawings show the preferred embodiments of this application, but do not limit the patent scope of this application. This application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure content of this application more thorough and comprehensive. Although this application has been described in detail with reference to the foregoing embodiments, for those skilled in the art, they can still modify the technical solutions recorded in the foregoing specific embodiments, or perform equivalent replacements on some of the technical features. Any equivalent structure directly or indirectly using the content of this application's specification and drawings in other related technical fields is equally within the scope of the patent protection of this application. The non-company software tools or components that appear in the embodiments of this application are only introduced by way of example and do not represent actual use.

Claims

1. A method for predicting purchase of target insurance, characterized in that: The steps include: From the user portrait information of the first insurance type, obtain the first type of factor data that influences the target user to purchase the second insurance type, wherein the first insurance type represents a non-target insurance type and the second insurance type represents a target insurance type; Analyze the first type of factor data to determine whether the target user has a second type of insurance subject matter; If the target user has an object of the second insurance type, obtaining the second type of factor data that influences the target user to purchase the second insurance type; Combining the first category factor data and the second category factor data to obtain the second insurance purchase prediction factor data; If the target user does not have the second insurance subject matter, the second insurance purchase prediction factor data will only be obtained based on the first type of factor data; Preprocessing the second insurance purchase prediction factor data to obtain target prediction data; The target prediction data is input into the trained second insurance type purchase prediction model, and the second insurance type purchase data is output.

2. The target insurance purchase prediction method according to claim 1, characterized in that: Before executing the step of acquiring the first type of factor data that influences the target user to purchase the second insurance type from the user portrait information of the first insurance type, the method further includes: Obtaining page browsing data and social sharing data of the target user according to a page browsing monitoring program of the first insurance business APP used by the target user; Based on the page browsing data and social sharing data, statistics are collected on the browsing time and sharing index of the target user for the subject matter corresponding to the second insurance type; The target user's relevant browsing time and sharing index for the subject matter corresponding to the second insurance type are used as supplementary information of the user portrait, and are cached in the user portrait information of the target user.

3. The target insurance purchase prediction method according to claim 2, characterized in that: The step of counting the relevant browsing time and sharing index of the target user for the subject matter corresponding to the second insurance type based on the page browsing data and social sharing data specifically includes: Analyze the page browsing data and the social sharing data using natural language processing technology to analyze browsing behaviors and social sharing behaviors related to the second insurance subject matter; Based on the browsing behavior and the social sharing behavior, statistics are collected on the target user's browsing time and sharing index for the subject matter corresponding to the second insurance type.

4. The target insurance purchase prediction method according to claim 2, characterized in that: After executing the step of using the target user's browsing time and sharing index for the subject matter corresponding to the second insurance type as user portrait supplementary information and caching it in the user portrait information of the target user, the method further includes: Analyzing the social relationship network of the target user using a graph neural network; According to the social relationship network analysis result, identifying the number of first insurance type users having second insurance type subject matter in the social relationship network of the target user, and the ratio of the number of first insurance type users having second insurance type subject matter in the social relationship network of the target user; The number of users of the first insurance type and the proportion value are used as supplementary information of the user portrait and are cached in the user portrait information of the target user.

5. The target insurance purchase prediction method according to any one of claims 1 to 4, characterized in that: The step of obtaining the first type of factor data that influences the target user to purchase the second insurance type from the user portrait information of the first insurance type specifically includes: Obtaining pre-set data tags that affect the purchase of the second insurance; Identify cache address information for storing user profile information of the first insurance type; Using the data tag and the identification information of the target user as the common search field and the cache library address information as the target search address, the cached data that affects the target user's purchase of the second insurance type is filtered out from the local business data cache library as the first type of factor data.

6. The target insurance purchase prediction method according to claim 5, characterized in that: The step of preprocessing the second insurance purchase prediction factor data to obtain target prediction data specifically includes: Performing deduplication processing, outlier processing, and data field format consistency processing on the second insurance purchase prediction factor data to obtain pre-processed second insurance purchase impact data; According to a preset vectorized representation dictionary, feature vectorization processing is performed on each of the impact data in the preprocessed second insurance purchase impact data to obtain a feature vectorization processing result; The feature vectorization processing result is set as the target prediction data.

7. The target insurance purchase prediction method according to claim 1, characterized in that: Before executing the step of inputting the target prediction data into the trained second insurance purchase prediction model and outputting the second insurance purchase data, the method further includes: From the target business information cache, select users who purchased both the first and second insurance types to generate a reference user group; Obtaining user portrait information corresponding to each user in the reference user group, and second insurance purchase data corresponding to each user; Analyze the user portrait information corresponding to each user in the reference user group to obtain the first type factor data and the second type factor data corresponding to each user; According to the user identification information, the corresponding first-category factor data and the second-category factor data are merged, and the second insurance purchase influence data corresponding to each user is obtained; According to a preset vectorized representation dictionary, feature vectorization processing is performed on each of the impact data in the second insurance purchase impact data corresponding to each user, to obtain a feature vectorization processing result; Based on the user identification information, the second insurance purchase data corresponding to different users and the feature vector quantization processing results are grouped and sorted to obtain the grouped sorting results as model training data, wherein each model training data includes the second insurance purchase data and purchase influence data corresponding to a user; The model training data is input into the second insurance purchase prediction model to be trained, the weight relationship between the second insurance purchase data and the purchase influence data is learned, and the second insurance purchase prediction model that has been learned and trained is obtained.

8. A target insurance purchase prediction device, characterized in that: include: A first-category factor data acquisition module, used to acquire first-category factor data that influences a target user to purchase a second insurance type from user portrait information of a first insurance type, wherein the first insurance type represents a non-target insurance type and the second insurance type represents a target insurance type; A first-category factor data analysis and judgment module, used to analyze the first-category factor data to determine whether the target user has a second-category insurance subject matter; The first branch processing module is used for obtaining the second type of factor data that affects the target user to purchase the second insurance type if the target user has the second insurance type subject matter; A factor data merging module, used to merge the first type of factor data and the second type of factor data to obtain the second insurance purchase prediction factor data; A second branch processing module is used for obtaining the second insurance purchase prediction factor data only according to the first type of factor data if the target user does not have the second insurance subject matter; A target prediction data acquisition module, used for preprocessing the second insurance purchase prediction factor data to obtain target prediction data; The second insurance type purchase prediction module is used to input the target prediction data into the trained second insurance type purchase prediction model and output the second insurance type purchase data.

9. A computer device, characterized in that: It comprises a memory and a processor, wherein the memory stores computer-readable instructions, and when the processor executes the computer-readable instructions, the steps of the target insurance purchase prediction method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the steps of the target insurance purchase prediction method according to any one of claims 1 to 7 are implemented.

Images