Information pushing method, device and equipment
By building user profiles and optimizing the information push process, the banking system can provide users with personalized information pushes, solving the problem of low information matching and improving user satisfaction and service targeting.
Patent Information
- Application Number
- CN202610313041.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-03-13
- Publication Date
- 2026-06-26
AI Technical Summary
When pushing information to users, the banking system fails to fully consider multiple factors such as users' interests and preferences, resulting in a low degree of matching between the information and users' actual needs, and failing to meet users' personalized needs.
By acquiring multi-source data from target users, performing feature extraction and clustering, constructing user profile tags, using a large language model for semantic analysis, combining content generation models to generate personalized information, and sending the information through optimized information push channels.
It achieves a deep understanding of user needs and preferences, provides customized push content, improves service targeting and user satisfaction, and enhances the flexibility and timeliness of push content.
Smart Images

Figure CN122293727A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of data processing technology, and in particular to information push methods, apparatus and equipment. Background Technology
[0002] With the development of information technology, banks have accumulated massive amounts of user data, including basic information, transaction records, browsing behavior, etc. In today's increasingly competitive banking environment, banks need to provide users with more accurate and personalized services to improve user satisfaction and loyalty.
[0003] In related technologies, when banking systems push information to users, banks do not conduct sufficient data mining on users and only push object information based on simple transaction data. They do not fully consider multi-dimensional factors such as users' interests and preferences, resulting in a low degree of matching between the pushed information and users' actual needs, making it difficult to meet users' personalized needs.
[0004] Therefore, when banking systems push information to users in related technologies, there is a problem that the pushed information cannot meet the personalized needs of users. Summary of the Invention
[0005] The purpose of this application is to provide an information push method, apparatus, and device to solve the problem that the information pushed by the banking system to users cannot meet the personalized needs of users.
[0006] Firstly, this application provides an information push method, the method comprising: Acquire multi-source data of the target user; the multi-source data includes user basic data, user transaction data, and browsing behavior data; Feature extraction is performed on the multi-source data to obtain the user profile tags of the target user; A large language model is used to perform semantic analysis on the user profile tags to obtain the user profile analysis results of the target user. The user profile analysis results are processed using a content generation model to generate push information for the target user; Based on the user profile analysis results and the information to be pushed, a target information push channel is determined, and the information to be pushed is sent to the target user using the target information push channel.
[0007] In one possible implementation, the step of extracting features from the multi-source data to obtain the user profile tags of the target user includes: Feature extraction is performed on the multi-source data to obtain user profile features; the user profile features include user lifecycle features, user transaction features, and user browsing features. Clustering algorithms are used to classify the user lifecycle features to determine the lifecycle stage sub-labels of the target user; Clustering algorithms are used to classify the user transaction features and user browsing features to determine the user interest sub-tags and risk preference sub-tags of the target user; The lifecycle stage sub-tags, user interest sub-tags, and risk preference sub-tags are used as user profile tags for the target user.
[0008] In one possible implementation, the step of performing semantic analysis on the user profile tags using a large language model to obtain the user profile analysis results of the target user includes: For each sub-tag in the user profile tags, perform the following operations: The sub-tags are semantically expanded using a large language model to obtain a user profile language. The user profile language is processed using a large language model to obtain a user profile semantic vector; The semantic vectors of multiple user profiles corresponding to multiple sub-tags are used as the user profile analysis results of the target user.
[0009] In one possible implementation, the step of processing the user profile analysis results using a content generation model to generate push information for the target user includes: The user profile analysis results and real-time object data are input into the content generation model to determine the target audience and generate push text. The feature information of the object to be pushed and the user interest sub-tags are input into the content generation model to generate a push image; The push text and the push image are used as the push information to be pushed to the target user.
[0010] In one possible implementation, determining the target information push channel based on the user profile analysis results and the information to be pushed includes: Based on the channel preference information in the user profile analysis results, the first information push channel is determined; Based on the importance of the target object in the information to be pushed, a second information push channel corresponding to the importance level is determined according to a table showing the relationship between importance level and information push channel. If it is determined that the first information push channel and the second information push channel are the same channel, then the first information push channel or the second information push channel shall be used as the target information push channel. If it is determined that the first information push channel and the second information push channel are different channels, then the first information push channel and the second information push channel will be used as the target information push channel.
[0011] In one possible implementation, the method further includes, before sending the push notification to the target user: Based on the user behavior data in the multi-source data, a time series analysis algorithm is used to determine the time point when the target user received the information. The user level of the target user is determined based on the user profile tags, and the information push strategy corresponding to the user level is determined based on the relationship table between the user level and the information push strategy.
[0012] In one possible implementation, sending the information to be pushed to the target user using the target information push channel includes: At the time point when the information is received, the information push strategy is used to send the information to be pushed to the target user.
[0013] Secondly, this application provides an information push device, the device comprising: The data acquisition module is configured to acquire multi-source data of the target user; the multi-source data includes user basic data, user transaction data, and browsing behavior data. The user profile tag determination module is configured to extract features from the multi-source data to obtain the user profile tags of the target user. The semantic analysis module is configured to perform semantic analysis on the user profile tags using a large language model to obtain the user profile analysis results of the target user. The push notification generation module is configured to process the user profile analysis results using a content generation model to generate push notifications for the target user. The information push channel determination module is configured to determine the target information push channel based on the user profile analysis results and the information to be pushed, and to send the information to be pushed to the target user using the target information push channel.
[0014] In one possible implementation, the feature extraction of the multi-source data to obtain the user profile tags of the target user, wherein the user profile tag determination module is specifically configured as follows: Feature extraction is performed on the multi-source data to obtain user profile features; the user profile features include user lifecycle features, user transaction features, and user browsing features. Clustering algorithms are used to classify the user lifecycle features to determine the lifecycle stage sub-labels of the target user; Clustering algorithms are used to classify the user transaction features and user browsing features to determine the user interest sub-tags and risk preference sub-tags of the target user; The lifecycle stage sub-tags, user interest sub-tags, and risk preference sub-tags are used as user profile tags for the target user.
[0015] In one possible implementation, the semantic analysis module is specifically configured to perform semantic analysis on the user profile tags using a large language model to obtain the user profile analysis results of the target user. For each sub-tag in the user profile tags, perform the following operations: The sub-tags are semantically expanded using a large language model to obtain a user profile language. The user profile language is processed using a large language model to obtain a user profile semantic vector; The semantic vectors of multiple user profiles corresponding to multiple sub-tags are used as the user profile analysis results of the target user.
[0016] In one possible implementation, the user profile analysis results are processed using a content generation model to generate push notification information for the target user. Specifically, the push notification information generation module is configured as follows: The user profile analysis results and real-time object data are input into the content generation model to determine the target audience and generate push text. The feature information of the object to be pushed and the user interest sub-tags are input into the content generation model to generate a push image; The push text and the push image are used as the push information to be pushed to the target user.
[0017] In one possible implementation, the step of determining the target information push channel based on the user profile analysis results and the information to be pushed, wherein the information push channel determination module is specifically configured as follows: Based on the channel preference information in the user profile analysis results, the first information push channel is determined; Based on the importance of the target object in the information to be pushed, a second information push channel corresponding to the importance level is determined according to a table showing the relationship between importance level and information push channel. If it is determined that the first information push channel and the second information push channel are the same channel, then the first information push channel or the second information push channel shall be used as the target information push channel. If it is determined that the first information push channel and the second information push channel are different channels, then the first information push channel and the second information push channel will be used as the target information push channel.
[0018] In one possible implementation, before sending the push notification to the target user, the device further includes: The information reception time point determination module is configured to determine the information reception time point of the target user based on user behavior data in the multi-source data and using a time series analysis algorithm. The information push strategy determination module is configured to determine the user level of the target user based on the user profile tags, and to determine the information push strategy corresponding to the user level based on the relationship lookup table between user level and information push strategy.
[0019] In one possible implementation, the step of sending the information to be pushed to the target user using the target information push channel is specifically configured as follows: At the time point when the information is received, the information push strategy is used to send the information to be pushed to the target user.
[0020] Thirdly, this application provides an electronic device, comprising: Processor and memory; The memory is used to store the processor-executable instructions; The processor is configured to execute the instructions to implement the information push method as provided in any of the first aspects of this application.
[0021] Fourthly, this application provides a computer-readable storage medium that, when the instructions in the computer-readable storage medium are executed by a processor of an electronic device, enables the electronic device to perform the information push method as described in any of the first aspects of this application.
[0022] Fifthly, this application provides a computer program product, including a computer program that, when executed by a processor of an electronic device, enables the electronic device to perform an information push method as described in any of the first aspects of this application.
[0023] The technical solutions provided by the embodiments of this application bring at least the following beneficial effects: In this embodiment of the application, by deeply constructing user profiles, a comprehensive understanding of user needs and preferences is achieved, enabling banks to provide customized push content for each user, significantly improving service targeting and user satisfaction; the push content dynamically generated by the large model is more flexible and novel, and by combining real-time information and user characteristics, the information to be pushed is more timely and attractive.
[0024] It should be understood that the above general description and the following detailed description are merely exemplary and explanatory, and do not limit this application. Based on common knowledge in the art, the above preferred conditions can be combined arbitrarily to obtain various preferred embodiments of this application. Attached Figure Description
[0025] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the embodiments of this application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0026] Figure 1 This is a schematic diagram illustrating an application scenario of an information push method provided in an embodiment of this application; Figure 2 This is a schematic diagram of the overall process of an information push method provided in an embodiment of this application; Figure 3 A schematic diagram of the process for obtaining the user profile tags of the target user in step 202 of this application embodiment; Figure 4 A flowchart illustrating the process of determining the user profile analysis results of the target user in step 203 of this application embodiment; Figure 5 This is a flowchart illustrating the process of generating the target user's push information in step 204, as provided in an embodiment of this application. Figure 6 A flowchart illustrating the process of determining the target information push channel in step 205 of this application embodiment; Figure 7 This is a schematic diagram of the structure of an information push device provided in an embodiment of this application; Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation
[0027] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. The described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.
[0028] Furthermore, in the description of the embodiments of this application, unless otherwise stated, " / " means "or". For example, A / B can mean A or B. The "and / or" in the text is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, and B exists alone. In addition, in the description of the embodiments of this application, "multiple" means two or more.
[0029] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature.
[0030] With the development of information technology, banks have accumulated massive amounts of user data, including basic information, transaction records, browsing behavior, etc. In today's increasingly competitive banking environment, banks need to provide users with more accurate and personalized services to improve user satisfaction and loyalty.
[0031] In related technologies, when banking systems push information to users, banks do not conduct sufficient data mining on users and only push object information based on simple transaction data. They do not fully consider multi-dimensional factors such as users' interests and preferences, resulting in a low degree of matching between the pushed information and users' actual needs, making it difficult to meet users' personalized needs.
[0032] Therefore, when banking systems push information to users in related technologies, there is a problem that the pushed information cannot meet the personalized needs of users.
[0033] In view of this, this application provides an information push method, apparatus, and device to solve the problem that the information pushed by the banking system to users cannot meet the personalized needs of users.
[0034] The inventive concept of this application can be summarized as follows: First, multi-source data of the target user is obtained, and the multi-source data is processed to obtain user profile tags of the target user; then, a large language model is used to perform semantic analysis on the user profile tags to obtain the user profile analysis results of the target user; furthermore, a content generation model is used to process the user profile analysis results to generate push information for the target user; finally, the push information is sent to the target user through the target information push channel.
[0035] In summary, the embodiments of this application, by deeply constructing user profiles and comprehensively understanding user needs and preferences, enable banks to provide customized push content for each user, significantly improving service targeting and user satisfaction; the push content dynamically generated by the large model is more flexible and novel, and combined with real-time information and user characteristics, makes the information to be pushed more timely and attractive.
[0036] After introducing the main inventive concepts of the embodiments of this application, the following is a brief description of the application scenarios to which the technical solutions of the embodiments of this application are applicable. It should be noted that the application scenarios described below are only for illustrating the embodiments of this application and are not intended to limit the scope. In specific implementation, the technical solutions provided by the embodiments of this application can be flexibly applied according to actual needs.
[0037] For ease of understanding, the following detailed description of an information push method provided by an embodiment of this application, in conjunction with the accompanying drawings, will be provided: like Figure 1 The diagram illustrates an application scenario of an information push method provided in this embodiment. The diagram includes: a network 10, a server 20, and a storage device 30. The server 20 obtains multiple sources via the network, such as basic user data, user transaction data, and browsing behavior data. Using the method provided in this embodiment, it processes the multi-source data to deeply construct user profiles, comprehensively understanding user needs and preferences, enabling the bank to provide customized push content for each user.
[0038] The description in this application focuses on a single server only. However, those skilled in the art should understand that the illustrated network 10, server 20, and memory 30 are intended to illustrate the operation of the electronic devices, servers, and memory involved in the technical solutions of this application. The detailed description of a single server and memory is at least for ease of explanation and does not imply any limitation on the number, type, or location of servers. It should be noted that adding additional modules to or removing individual modules from the illustrated environment will not change the underlying concept of the exemplary embodiments of this application. Furthermore, although detailed descriptions are provided for ease of explanation... Figure 1 The diagram shows a bidirectional arrow from memory 30 to server 20, but those skilled in the art will understand that the sending and receiving of the aforementioned data also needs to be achieved through network 10.
[0039] It should be noted that the memory in the embodiments of this application can be, for example, a cache system, hard disk storage, memory storage, etc. Furthermore, the information push method proposed in this application is not only applicable to… Figure 1 The application scenarios shown can also be used in other possible application scenarios, and the embodiments of this application do not impose any limitations.
[0040] Based on the above description, the information push method provided in this application embodiment has the following overall process: Figure 2 As shown, it includes the following: In step 201, multi-source data of the target user is obtained. This multi-source data includes basic user data, user transaction data, and browsing behavior data.
[0041] It should be noted that this application embodiment obtains multi-source data on target users from the bank's internal core systems (such as user information management systems, transaction systems, and credit systems), online channels (mobile banking apps, online banking browsing logs), and third-party cooperative platforms (such as credit rating agencies and consumer data providers). This data includes basic user data such as age, gender, and occupation; user economic data such as asset size, income and expenditure, and debt status; user transaction data such as transaction frequency, amount, and product preferences; browsing behavior data such as viewed pages, dwell time, and search keywords; and user social data, including social media activity and areas of interest. The multi-source data is not limited to the data mentioned above. For example, this application embodiment also obtains user credit information through cooperation with third-party credit rating agencies to supplement the bank's internal credit assessment data.
[0042] In one possible implementation, this application embodiment will also perform data cleaning and preprocessing on the multi-source data, including the following: cleaning the collected multi-source data to remove duplicate, erroneous, and missing values; processing missing values and noisy data using methods such as data interpolation and smoothing; and standardizing data of different formats to prepare for subsequent analysis. For example, identifying and correcting abnormally large values in transaction amount data caused by system failures, and standardizing user birth dates of different formats to "YYYY-MM-DD" format.
[0043] In step 202, feature extraction is performed on the multi-source data to obtain user profile tags for the target user.
[0044] In one possible implementation, the process of obtaining the target user's user profile tags in step 202 is as follows: Figure 3 As shown, it can be implemented as follows: In step 301, feature extraction is performed on the multi-source data to obtain user profile features, which include user lifecycle features, user transaction features, and user browsing features.
[0045] In step 302, a clustering algorithm is used to classify user lifecycle features and determine the lifecycle stage sub-labels of the target user.
[0046] In step 303, a clustering algorithm is used to classify user transaction features and user browsing features to determine the user interest sub-labels and risk preference sub-labels of the target user.
[0047] In step 304, the lifecycle stage sub-tag, user interest sub-tag, and risk preference sub-tag are used as user profile tags for the target user.
[0048] This application uses Python's Scikit-learn library and the deep learning framework TensorFlow for feature extraction and user profile construction. It extracts user transaction features such as transaction amount, transaction frequency, and transaction product type from user transaction data; it extracts user browsing features such as page type, browsing duration, and search keywords from browsing behavior data; and it uses clustering algorithms (such as K-Means clustering) to classify user risk preferences and divide life cycle stage labels according to user life cycle stage-related features (such as age, marital status, and children).
[0049] For example, cluster analysis can be performed on the user lifecycle characteristics, user transaction characteristics, and user browsing characteristics of 1 million users. K-Means clustering can be used to label users with risk preference sub-tags such as conservative, stable, and aggressive, user interest sub-tags such as stock investment, fund preference, and insurance needs, and user lifecycle stages such as single, newlywed, parenting, and retirement.
[0050] It should be noted that the user profile tags for target users are not limited to the three sub-tags mentioned above.
[0051] The steps described above for building user profile tags can deeply construct user profiles, fully understand user needs and preferences, and enable banks to provide customized push information for each user, greatly improving service targeting and user satisfaction.
[0052] In step 203, a large language model is used to perform semantic analysis on the user profile tags to obtain the user profile analysis results of the target user.
[0053] In one possible implementation, the process for determining the user profile analysis results of the target user in step 203 is as follows: Figure 4 As shown, it can be implemented as follows: For each sub-tag in the user profile tag, perform the following operations: In step 401, the sub-tags are semantically expanded using a large language model to obtain the user profile language.
[0054] In step 402, the user profile language is processed using a large language model to obtain the user profile semantic vector.
[0055] In step 403, the semantic vectors of multiple user profiles corresponding to multiple sub-tags are used as the user profile analysis results of the target user.
[0056] This application's embodiments are based on the open-source BERT model, fine-tuned and trained to build a dedicated NLP engine for the banking sector. A large amount of text data, including bank product introduction documents, financial news, and user consultation records, is collected to pre-train the BERT model, enabling it to understand the semantics and grammatical rules of the banking sector. This engine is then used to semantically understand and expand user profile tags. Combined with knowledge graphs, user profile tags are transformed into understandable linguistic concepts. For example, the tag "high-risk preference" is expanded to semantic expressions such as "pursuing high returns and willing to bear higher investment risks," providing input for subsequent content generation models to generate push notifications.
[0057] For example, if the input risk preference sub-label is "aggressive" and the user interest sub-label is "tech stocks", the NLP engine can understand that the user has a demand for high-risk, high-return technology stock investment products and generate relevant semantic vectors for subsequent content generation models.
[0058] In step 204, the user profile analysis results are processed using a content generation model to generate push information for the target user.
[0059] In one possible implementation, step 204 uses a content generation model to process the user profile analysis results and generate push notifications for the target user. The process of this step is as follows: Figure 5 As shown, it can be implemented as follows: In step 501, the user profile analysis results and real-time object data input content are used to generate a model, determine the target audience for push notifications, and generate push notification text.
[0060] In step 502, the feature information of the object to be pushed and the user's interest sub-tags are input into the content generation model to generate the push image.
[0061] In step 503, push text and push images are selected as the information to be pushed to the target user.
[0062] This application uses a generative adversarial network (GAN) or a Transformer-based generative model to generate personalized push information based on user profile analysis results.
[0063] For example, for push notification text, by combining real-time market data, the text is generated that includes product advantages, expected returns, and risk warnings, while conforming to the user's language style and comprehension level. For instance, for a young user with an aggressive risk appetite and an interest in the technology industry, the following push notification text for an equity fund is generated: "Dear user, the technology industry is developing rapidly. This [fund name] focuses on high-quality stocks in cutting-edge technology fields. With the help of a professional fund manager team that accurately grasps market trends, the expected annualized return can reach [X]%. Of course, high returns come with certain risks, but with your adventurous spirit and keen insight into technology, we believe you can handle it well and start a new journey of wealth appreciation!" For push notification images, image generation technology based on generative adversarial networks (GANs) is used. The user's characteristic information (such as product type, target user group) and user interest sub-tags (such as liking technology elements, preferring a minimalist style) are input to generate push notification images (such as product covers) that meet the user's needs. For instance, a cover image combining technologically advanced charts and a simple, modern style is generated for an innovative financial product targeting young technology enthusiasts, while a cover image combining a warm family scene and elements representing wealth is generated for a stable financial product targeting elderly users.
[0064] In step 205, based on the user profile analysis results and the information to be pushed, the target information push channel is determined, and the information to be pushed is sent to the target user through the target information push channel.
[0065] In one possible implementation, step 205 determines the target information push channel based on the user profile analysis results and the information to be pushed. The process of this step is as follows: Figure 6 As shown, it can be implemented as follows: In step 601, the first information push channel is determined based on the channel preference information in the user profile analysis results.
[0066] In step 602, based on the importance of the target object in the information to be pushed, a second information push channel corresponding to the importance level is determined according to a table relating importance level to information push channels.
[0067] In step 603, if it is determined that the first information push channel and the second information push channel are the same channel, then the first information push channel or the second information push channel is used as the target information push channel.
[0068] In step 604, if it is determined that the first information push channel and the second information push channel are different channels, then the first information push channel and the second information push channel are used as the target information push channel.
[0069] This application integrates multiple user outreach channels for banks, including mobile banking app push notifications, SMS, email, and telemarketing. It analyzes the characteristics and applicable scenarios of each channel, establishes a channel effectiveness evaluation model, and selects the optimal combination of push channels based on channel preference information and the importance of the target audience in the user profile analysis results.
[0070] For example, younger users tend to prefer push notifications from mobile banking apps, while older users are more receptive to SMS messages. Important product information is pushed via a combination of SMS and app messages, while regular push notifications are sent via a single channel based on user habits.
[0071] In one possible implementation, before sending the information to be pushed to the target user, this application embodiment will also use time series analysis algorithms (such as ARIMA model) based on user behavior data in multi-source data to predict the probability of users using banking services in different time periods, and combine user life scenario information (such as working hours of office workers and holiday arrangements) to determine the best push time (i.e. the time point of receiving information).
[0072] For example, if analysis reveals that a working professional uses mobile banking most frequently between 8 and 9 pm from Monday to Friday and has a habit of traveling on weekends, then information on wealth management products will be pushed to them at 8 pm on weekdays, and information on travel consumer loan promotions will be pushed to them on Friday afternoons.
[0073] In one possible implementation, before sending the information to be pushed to the target user, this embodiment of the application will also determine the user level of the target user based on the user profile tags, and determine the information push strategy corresponding to the user level based on the relationship table between user level and information push strategy.
[0074] For example, users are categorized into different groups based on user profile tags, such as high-value users, potential users, and ordinary users. For high-value users, a dedicated service team is established to provide one-on-one recommendations for high-end financial products and customized investment solutions, through regular in-person visits and dedicated app push notifications. For potential users, information on promotional activities (such as opening an account with a bonus, or waived fees for first-time investments) and beginner financial knowledge pushes are used to attract them to further utilize the bank's services. For ordinary users, information on popular financial products is pushed based on their interests and preferences; for example, high-net-worth users receive a personalized investment analysis report every month, and newly registered potential users receive a basic financial knowledge tip every day for a week.
[0075] This application embodiment can send the information to be pushed to the target user at the time of information reception using an information push strategy. It fully considers the user's channel preferences, behavioral patterns and life scenarios, and realizes that the information will be pushed to the target user at the right time, through the right channel and in the right way. This improves the reach rate and click rate of the push information, reduces the cost of invalid pushes, and improves the bank's marketing efficiency and return on investment.
[0076] In summary, the embodiments of this application, by deeply constructing user profiles and comprehensively understanding user needs and preferences, enable banks to provide customized push content for each user, significantly improving service targeting and user satisfaction; the push content dynamically generated by the large model is more flexible and novel, and combined with real-time information and user characteristics, makes the information to be pushed more timely and attractive.
[0077] Based on the same inventive concept, embodiments of this application also provide an information push device, such as... Figure 7 As shown, the device 700 includes: The data acquisition module 701 is configured to acquire multi-source data of the target user; the multi-source data includes user basic data, user transaction data, and browsing behavior data. User profile tag determination module 702 is configured to extract features from the multi-source data to obtain user profile tags for the target user. The semantic analysis module 703 is configured to perform semantic analysis on the user profile tags using a large language model to obtain the user profile analysis results of the target user. The push notification generation module 704 is configured to process the user profile analysis results using a content generation model to generate push notifications for the target user. The information push channel determination module 705 is configured to determine the target information push channel based on the user profile analysis results and the information to be pushed, and to send the information to be pushed to the target user using the target information push channel.
[0078] In one possible implementation, the feature extraction of the multi-source data to obtain the user profile tags of the target user, wherein the user profile tag determination module is specifically configured as follows: Feature extraction is performed on the multi-source data to obtain user profile features; the user profile features include user lifecycle features, user transaction features, and user browsing features. Clustering algorithms are used to classify the user lifecycle features to determine the lifecycle stage sub-labels of the target user; Clustering algorithms are used to classify the user transaction features and user browsing features to determine the user interest sub-tags and risk preference sub-tags of the target user; The lifecycle stage sub-tags, user interest sub-tags, and risk preference sub-tags are used as user profile tags for the target user.
[0079] In one possible implementation, the semantic analysis module is specifically configured to perform semantic analysis on the user profile tags using a large language model to obtain the user profile analysis results of the target user. For each sub-tag in the user profile tags, perform the following operations: The sub-tags are semantically expanded using a large language model to obtain a user profile language. The user profile language is processed using a large language model to obtain a user profile semantic vector; The semantic vectors of multiple user profiles corresponding to multiple sub-tags are used as the user profile analysis results of the target user.
[0080] In one possible implementation, the user profile analysis results are processed using a content generation model to generate push notification information for the target user. Specifically, the push notification information generation module is configured as follows: The user profile analysis results and real-time object data are input into the content generation model to determine the target audience and generate push text. The feature information of the object to be pushed and the user interest sub-tags are input into the content generation model to generate a push image; The push text and the push image are used as the push information to be pushed to the target user.
[0081] In one possible implementation, the step of determining the target information push channel based on the user profile analysis results and the information to be pushed, wherein the information push channel determination module is specifically configured as follows: Based on the channel preference information in the user profile analysis results, the first information push channel is determined; Based on the importance of the target object in the information to be pushed, a second information push channel corresponding to the importance level is determined according to a table showing the relationship between importance level and information push channel. If it is determined that the first information push channel and the second information push channel are the same channel, then the first information push channel or the second information push channel shall be used as the target information push channel. If it is determined that the first information push channel and the second information push channel are different channels, then the first information push channel and the second information push channel will be used as the target information push channel.
[0082] In one possible implementation, before sending the push notification to the target user, the device further includes: The information reception time point determination module is configured to determine the information reception time point of the target user based on user behavior data in the multi-source data and using a time series analysis algorithm. The information push strategy determination module is configured to determine the user level of the target user based on the user profile tags, and to determine the information push strategy corresponding to the user level based on the relationship lookup table between user level and information push strategy.
[0083] In one possible implementation, the step of sending the information to be pushed to the target user using the target information push channel is specifically configured as follows: At the time point when the information is received, the information push strategy is used to send the information to be pushed to the target user.
[0084] The following reference Figure 8 To describe an electronic device 130 according to this embodiment of the present application. Figure 8 The electronic device 130 shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.
[0085] like Figure 8 As shown, the electronic device 130 is presented in the form of a general electronic device. The components of the electronic device 130 may include, but are not limited to: at least one processor 131, at least one memory 132, and a bus 133 connecting different system components (including memory 132 and processor 131).
[0086] Bus 133 represents one or more of several bus structures, including a memory bus or memory controller, peripheral bus, processor, or local bus using any of the various bus structures.
[0087] The memory 132 may include a readable medium in the form of volatile memory, such as random access memory (RAM) 1321 and / or cache memory 1322, and may further include read-only memory (ROM) 1323.
[0088] The memory 132 may also include a program / utility 1325 having a set (at least one) of program modules 1324, including but not limited to: an operating system, one or more application programs, other program modules, and program data, each or some combination of these examples may include an implementation of a network environment.
[0089] Electronic device 130 can also communicate with one or more external devices 134 (e.g., keyboard, pointing device, etc.), and with one or more devices that enable a user to interact with electronic device 130, and / or with any device that enables electronic device 130 to communicate with one or more other electronic devices (e.g., router, modem, etc.). This communication can be performed via input / output (I / O) interface 135. Furthermore, electronic device 130 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 136. As shown, network adapter 136 communicates with other modules used in electronic device 130 via bus 133. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with electronic device 130, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.
[0090] In an exemplary embodiment, this application also provides a computer-readable storage medium including instructions, such as a memory 132 including instructions, which can be executed by a processor 131 of an electronic device 130 to complete the aforementioned information push method. Optionally, the computer-readable storage medium can be a non-transitory computer-readable storage medium, such as a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device.
[0091] In an exemplary embodiment, a computer program product is also provided, including a computer program that, when executed by a processor 131, implements the information push method provided in this application.
[0092] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0093] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0094] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0095] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0096] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.
Claims
1. An information push method, characterized in that, The method includes: Acquire multi-source data of the target user; the multi-source data includes user basic data, user transaction data, and browsing behavior data; Feature extraction is performed on the multi-source data to obtain the user profile tags of the target user; A large language model is used to perform semantic analysis on the user profile tags to obtain the user profile analysis results of the target user. The user profile analysis results are processed using a content generation model to generate push information for the target user; Based on the user profile analysis results and the information to be pushed, a target information push channel is determined, and the information to be pushed is sent to the target user using the target information push channel.
2. The method according to claim 1, characterized in that, The step of extracting features from the multi-source data to obtain the user profile tags for the target user includes: Feature extraction is performed on the multi-source data to obtain user profile features; the user profile features include user lifecycle features, user transaction features, and user browsing features. Clustering algorithms are used to classify the user lifecycle features to determine the lifecycle stage sub-labels of the target user; Clustering algorithms are used to classify the user transaction features and user browsing features to determine the user interest sub-tags and risk preference sub-tags of the target user; The lifecycle stage sub-tags, user interest sub-tags, and risk preference sub-tags are used as user profile tags for the target user.
3. The method according to claim 2, characterized in that, The step of using a large language model to perform semantic analysis on the user profile tags to obtain the user profile analysis results of the target user includes: For each sub-tag in the user profile tags, perform the following operations: The sub-tags are semantically expanded using a large language model to obtain a user profile language. The user profile language is processed using a large language model to obtain a user profile semantic vector; The semantic vectors of multiple user profiles corresponding to multiple sub-tags are used as the user profile analysis results of the target user.
4. The method according to claim 3, characterized in that, The step of processing the user profile analysis results using a content generation model to generate push notification information for the target user includes: The user profile analysis results and real-time object data are input into the content generation model to determine the target audience and generate push text. The feature information of the object to be pushed and the user interest sub-tags are input into the content generation model to generate a push image; The push text and the push image are used as the push information to be pushed to the target user.
5. The method according to claim 4, characterized in that, The step of determining the target information push channel based on the user profile analysis results and the information to be pushed includes: Based on the channel preference information in the user profile analysis results, the first information push channel is determined; Based on the importance of the target object in the information to be pushed, a second information push channel corresponding to the importance level is determined according to a table showing the relationship between importance level and information push channel. If it is determined that the first information push channel and the second information push channel are the same channel, then the first information push channel or the second information push channel shall be used as the target information push channel. If it is determined that the first information push channel and the second information push channel are different channels, then the first information push channel and the second information push channel will be used as the target information push channel.
6. The method according to claim 1, characterized in that, Before sending the information to be pushed to the target user, the method further includes: Based on the user behavior data in the multi-source data, a time series analysis algorithm is used to determine the time point when the target user received the information. The user level of the target user is determined based on the user profile tags, and the information push strategy corresponding to the user level is determined based on the relationship table between the user level and the information push strategy.
7. The method according to claim 6, characterized in that, Sending the information to be pushed to the target user using the target information push channel includes: At the time point when the information is received, the information push strategy is used to send the information to be pushed to the target user.
8. An information push device, characterized in that, The device includes: The data acquisition module is configured to acquire multi-source data of the target user; the multi-source data includes user basic data, user transaction data, and browsing behavior data. The user profile tag determination module is configured to extract features from the multi-source data to obtain the user profile tags of the target user. The semantic analysis module is configured to perform semantic analysis on the user profile tags using a large language model to obtain the user profile analysis results of the target user. The push notification generation module is configured to process the user profile analysis results using a content generation model to generate push notifications for the target user. The information push channel determination module is configured to determine the target information push channel based on the user profile analysis results and the information to be pushed, and to send the information to be pushed to the target user using the target information push channel.
9. An electronic device, characterized in that, include: Processor and memory; The memory is used to store the processor-executable instructions; The processor is configured to execute the instructions to implement the information push method as described in any one of claims 1-7.
10. A computer-readable storage medium, characterized in that, When the instructions in the computer-readable storage medium are executed by the processor of the electronic device, the electronic device is able to perform the information push method as described in any one of claims 1-7.
