Information pushing method and device, electronic equipment and medium
By identifying user tagging information in private traffic and selecting target topics in public traffic, the accuracy of information push is solved, ensuring that information is received by interested users and improving marketing effectiveness.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- GREAT WALL MOTOR CO LTD
- Filing Date
- 2023-09-27
- Publication Date
- 2026-08-04
AI Technical Summary
Current information push technologies lack accuracy, resulting in push information failing to achieve the expected results and even causing user resentment.
Based on user behavior data in private domain traffic, user tag information is determined, and candidate topics are determined by combining public domain traffic. Target topics are selected through similarity analysis, and a user list is generated for precise push notifications.
It achieves high accuracy in information delivery, ensuring that target topics are received by interested users, thereby maximizing marketing objectives.
Smart Images
Figure CN117312663B_ABST
Abstract
Description
Technical Field
[0001] This application belongs to the field of vehicle technology, and in particular relates to an information push method, device, electronic device and medium. Background Technology
[0002] Push notifications can help businesses discover potential users. In marketing operations, it's often necessary to push promotional information to target users. However, inaccurate push notifications may fail to achieve the desired results and could even easily cause user resentment.
[0003] Currently, in marketing operations, the theme of an event is typically determined by planners based on their personal experience and subjective judgment, and then related push notifications are sent to users. This method of information delivery often results in a disconnect between the planner's personal judgment and user needs, leading to uninteresting information and failing to achieve the desired promotional effect.
[0004] For example, some users are interested in vehicles with off-road capabilities, while others are not. When marketing and promoting a vehicle, if information emphasizing its off-road capabilities is pushed to users who are not interested in such features, it will inevitably cause resentment. Summary of the Invention
[0005] In view of this, embodiments of this application provide an information push method, apparatus, electronic device, and medium to improve the accuracy of information push.
[0006] The first aspect of this application provides an information push method, including:
[0007] Based on user behavior data in private domain traffic, the tag information of each user is determined. The tag information includes one or more feature values corresponding to private domain topics. The feature values are used to characterize the user's tendency towards the corresponding private domain topic.
[0008] Based on public domain traffic, several alternative topics to be pushed are determined;
[0009] Based on one or more of the private domain topics, a target topic is determined from a plurality of candidate topics. The target topic has a corresponding target private domain topic. The target private domain topic is a private domain topic among one or more private domain topics whose similarity to the target topic is greater than a preset threshold.
[0010] Based on the feature value of the target private domain topic for each user, a user list corresponding to the target topic is generated;
[0011] Based on the user list, the target topic is pushed to users.
[0012] A second aspect of this application provides an information push device, including:
[0013] The tag information determination module is used to determine the tag information of each user based on user behavior data in private domain traffic. The tag information includes one or more private domain topics and feature values of each private domain topic. The feature values are used to characterize the user's tendency towards the corresponding private domain topic.
[0014] The alternative topic determination module is used to determine multiple alternative topics to be pushed based on public domain traffic;
[0015] The target topic determination module is used to determine a target topic from multiple candidate topics based on one or more of the private domain topics. The target topic has a corresponding target private domain topic, and the target private domain topic is a private domain topic among one or more private domain topics whose similarity to the target topic is greater than a preset threshold.
[0016] The user list generation module is used to generate a user list corresponding to the target topic based on the feature value of the target private domain topic for each user.
[0017] The target topic push module is used to push the target topic based on the user list.
[0018] A third aspect of this application provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method described in the first aspect above.
[0019] A fourth aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in the first aspect above.
[0020] A fifth aspect of this application provides a computer program product that, when run on a server, causes the server to execute the method described in the first aspect.
[0021] Compared with the prior art, the embodiments of this application have the following advantages:
[0022] When using the method provided in this application for information push, on the one hand, the server can determine user tag information based on user behavior data in private domain traffic. The feature values included in these tag information can reflect the user's inclination towards various private domain topics. On the other hand, the server can determine candidate topics based on public domain traffic. Through the analysis of private domain topics and candidate topics, the server can determine target topics that are of interest to the public and private domain users. Based on users' inclination towards various private domain topics, a user list corresponding to the target topic can be generated. In this way, the server can push information to the target topic according to the user list corresponding to the target topic. The target topics determined in this application embodiment are topics that have popularity in public domain traffic and are of interest to private domain users in private domain traffic. Therefore, pushing information based on target topics can not only push topics with high popularity, but also push target topics to users who are interested in them, ensuring accurate information push and maximizing the achievement of marketing objectives. Attached Figure Description
[0023] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.
[0024] Figure 1 This is a schematic diagram of an information push method provided in an embodiment of this application;
[0025] Figure 2 This is a flowchart illustrating an information push method provided in an embodiment of this application;
[0026] Figure 3 This is a flowchart illustrating the steps of an information push method provided in an embodiment of this application;
[0027] Figure 4 This is a schematic diagram of user classification provided in an embodiment of this application;
[0028] Figure 5 This is a schematic diagram of an information push device provided in an embodiment of this application;
[0029] Figure 6 This is a schematic diagram of a server provided in an embodiment of this application. Detailed Implementation
[0030] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.
[0031] The technical solution of this application will be described below through specific embodiments.
[0032] When pushing information, to make the information interesting to users, it can be combined with current trending topics. For more targeted information delivery, relevant information can be pushed to users who are interested in and have a positive attitude towards related topics. Based on this, this application provides an information push method to improve the accuracy of information delivery.
[0033] Figure 1 This is a schematic diagram of an information push provided in an embodiment of this application. Figure 1 Private domain behavior in this context refers to user behavior data within private domain traffic. Private domain traffic can be traffic acquired through the information pusher's own digital platforms, social media accounts, and other channels, which can be freely controlled and reused. For example, automakers may have their own official accounts, websites, Weibo accounts, and online stores; the user traffic of these accounts constitutes the automaker's private domain traffic. Users within private domain traffic are those with a certain demand for related products or services. Targeting these users with information makes precise marketing easier. Therefore, in this embodiment, private domain behavior data can be collected from private domain traffic, cleaned, and then used for user interest analysis to determine the interests of users within the private domain traffic.
[0034] Public domain topics can be collected from public domain traffic, which refers to traffic from public platforms. For example, traffic from e-commerce platforms, public social media platforms, and short video platforms can be considered public domain traffic. Public domain topics are generally current trending topics and topics of interest to the general public. Users in private domain traffic are also a part of the general public, so these same public domain topics may also be of interest to them. Therefore, collecting public domain topics from public domain traffic to generate event topics can reduce reliance on the planner's personal experience and subjective judgment, allowing the selected topics to better match user needs.
[0035] After analyzing user interests based on private domain behavior data in private domain traffic and generating activity topics based on public domain topics in public domain traffic, we can predict user activity interests based on user interests and activity topics, thereby segmenting users and pushing matching information to segmented groups.
[0036] Figure 2 This is a schematic diagram of an information push process provided in an embodiment of this application. Combined with... Figure 2 The following is an example of a car manufacturer pushing vehicle information to users:
[0037] Figure 2 The proprietary platforms within the system can be the automaker's official WeChat account, website, Weibo account, online store, and other private platforms. Users can access these platforms and publish information. For example, users can browse, like, comment, and share content on the automaker's official WeChat account, website, Weibo account, and online store, thereby generating user data within the automaker's private traffic.
[0038] Automakers can embed tracking points on their own platforms to collect data. For example, they can collect data on user visits, likes, comments, and shares. This collected data can be automatically tagged, essentially determining user preferences for specific private domain topics based on user behavior data. For instance, an automaker's official WeChat account might include an article about the vehicle's music function. User visits, likes, comments, and shares of this article can reflect their interest and dislike for this feature. Analyzing each user's behavior data reveals their specific interest and dislike for the music function. This information can then be included in the user's tagging information. Since private domain traffic can encompass numerous private domain topics, data analysis can reveal each user's interest and dislike for multiple topics. This information can then be added to the user's tagging information. Based on this tagging information, sentiment analysis can be performed to determine user interests and emotional states. Based on each user's tag information, users can be segmented into groups that share similar interests and preferences on the same topic.
[0039] Figure 2 The entire internet can include public platforms, such as social media platforms, short video platforms, and vehicle-related vertical websites or forums. The public can publish information across this entire network. Automakers can use cloud services to perform Python web scraping on public platforms to obtain vehicle-related data from public traffic. This data can then be consolidated to extract public domain topics. These extracted topics can be scored based on their popularity, and high-scoring topics can be added to an inspiration pool. Topics in the inspiration pool can improve the selection of topics when automakers are planning their operational strategies.
[0040] By matching topics and user groups in the inspiration pool, we can match popular topics with corresponding subgroups. Then, based on the matching results, we can push corresponding event plans to users, so that users belonging to different user groups can receive different planned events.
[0041] Let User A illustrate the solution in this application. Assume that private domain topics may include music topics, off-road topics, and voice control topics.
[0042] User A can publish information on their own platform via their user terminal, and the automaker can collect user A's behavior data through cloud services. Based on this user behavior data, data analysis can be performed to determine user A's level of interest and dislike regarding music, off-road, and voice control topics. Based on this level of interest and dislike, user A can be categorized. For example, if user A likes and is interested in the vehicle's off-road capabilities but dislikes and is not interested in voice control, and likes the vehicle's music features but is not interested in vehicle-related music topics, then user A could be categorized as a potential user for music topics, a key seed user for off-road topics, or an unreachable user for voice control topics.
[0043] If off-roading is a popular topic in vehicle-related forums, it can be added to the inspiration pool. Since this off-roading topic is also included in private domain topics, planners can access it from the inspiration pool to create relevant content and generate different push notifications for different user categories. When pushing information about off-roading, since user A is a key seed user for this topic, the push notification intended for key seed users can be sent to user A.
[0044] If current trending topics also include music topics, these can be added to the inspiration pool. Since the music topic is also included in private domain topics, planners can retrieve it from the inspiration pool for relevant planning, thereby generating different push notifications for different user categories. When pushing information about a music topic, since user A is a potential user of that topic, the push notification intended for potential users can be sent to user A.
[0045] The above explanation uses a single user as an example. Generally, when pushing information, it's possible to plan based on multiple topics, generating push notifications corresponding to different topics. These notifications can be sent to key seed users. For example, if user A is a key seed user for the off-road topic, then user A can receive push notifications about off-road related vehicles; if user B is a key seed user for the music topic, then user B can receive push notifications about vehicles with better music features; if user C is a key seed user for the voice control topic, then user B can receive push notifications about vehicles with voice control functionality. Based on the method in this application, automakers can push different notifications to different users according to their interests and preferences, thereby improving the accuracy of information delivery.
[0046] The steps of the information push method in this application are described below with reference to specific embodiments.
[0047] Reference Figure 3 The diagram illustrates a flowchart of an information push method provided in an embodiment of this application, which may specifically include the following steps:
[0048] S301, Based on user behavior data in private domain traffic, determine the tag information of each user. The tag information includes feature values corresponding to one or more private domain topics. The feature values are used to characterize the user's tendency towards the corresponding private domain topic.
[0049] The execution subject of this embodiment can be a server, which can be a desktop computer, laptop, handheld computer, or cloud server, etc. This application embodiment does not limit the specific type of server.
[0050] User behavior data can include data generated from various actions performed by users on private platforms such as applications or websites. Examples include user access data, browsing data, likes data, comment data, and posting data.
[0051] The aforementioned user behavior data can be collected through event tracking. Event tracking refers to attaching data collection program code to the functional program code at the "operation node" where data needs to be collected, in order to capture, process, and send user behavior or events at the operation node. For example, data collection code can be attached to the program code for accessing, liking, commenting, and forwarding, thereby collecting user behavior data such as access, comments, likes, comments, and forwarding. For instance, automakers can use servers to collect event tracking data through their own brand's applications, official websites, customer service, and other channels, thereby obtaining data on the number of user comments, posts, likes, and shares about vehicles on multiple channels, as well as user access frequency data on websites, thus enabling analysis of vehicle-related topics.
[0052] After collecting user behavior data, data cleaning can be performed to remove missing or invalid values. Following data cleaning, data analysis can be conducted to determine the tagging information for each user.
[0053] When conducting data analysis, private domain topics can be extracted from user behavior data to establish a private domain topic library. Private domain topics can be those topics of interest to users within private domain traffic. In one possible implementation, text analysis can be performed on user behavior data to obtain private domain topics. Specifically, Natural Language Understanding (NLU) technology can be used for information extraction and intent recognition to form structured data. Then, K-means clustering algorithm is used to extract user behavior data related to vehicle model, function, and topic for correlation classification, thereby extracting private domain topics from the user behavior data. Private domain topics can include information related to vehicle model, function, and topic. For example, the extracted private domain topic could be the automatic opening of the door control function of vehicle model A on rainy days. In other embodiments of this application, private domain topics are summarized in simple language for ease of explanation.
[0054] The server can periodically collect user behavior data, thereby automatically extracting private domain topics to expand the topic tag library in terms of both quantity and category. In this embodiment, the acquisition of private domain topics can be automated and performed periodically, reducing manual labor and improving work efficiency.
[0055] Based on the extracted private domain topics, tag information for each user can be obtained. User tag information can include feature values corresponding to each private domain topic, which can be used to indicate the user's inclination towards that topic. A user's inclination towards a private domain topic can include interest inclination and emotional inclination. Interest inclination can be used to describe whether a user is interested in the private domain topic, while emotional inclination can be used to describe the user's likes or dislikes towards the topic. In this embodiment, a first feature value can be used to represent interest inclination, and a second feature value can be used to represent emotional inclination.
[0056] A user's interest in a private domain topic can be reflected by the frequency of their visits and interaction with that topic. Generally, users will visit and interact with topics they are interested in. Therefore, we can obtain user behavior data on interaction and frequency of visits for each private domain topic from user behavior data, and then calculate a first feature value based on this data. For example, we can assign corresponding interest scores to the frequency of visits and interactions; for instance, one point can be added for each visit, 0.5 points for each like, 1 point for each comment, 0.5 points for each share, and 3 points for each "not interested" click. When analyzing a user's interest in a private domain topic, we can extract the user's relevant visit frequency and interaction data for that topic from user behavior data, calculate the user's interest score for that topic based on this data, and then use this interest score as the first feature value to characterize the user's interest in the private domain topic.
[0057] Sentiment analysis algorithms are used to analyze user behavior data, determining the second characteristic value of each user's attitude towards each private domain topic. The server can pre-train a sentiment analysis algorithm or directly acquire pre-trained algorithms from other devices. This algorithm analyzes the topic tags associated with users, extracting their attitudes, evaluations, opinions, and sentiment tendencies towards different topic tags, and calculating sentiment scores. For example, the sentiment score range can be -100 to 100. A sentiment score between 40 and 100 indicates a positive attitude towards the private domain topic; a score between -40 and 40 indicates a neutral attitude; and a score between -40 and -100 indicates a negative attitude.
[0058] In one possible implementation, the two-dimensional vector composed of the first and second eigenvalues can be directly used as the aforementioned eigenvalues.
[0059] In another possible implementation, a user's user type on a private topic can be determined based on their first and second feature values. Then, a corresponding tag value can be determined based on this user type. For example, the tag values for key seed users, potential users, biased users, and unreached users are 4, 3, 2, and 1, respectively. The tag value can be used to characterize the user's inclination towards that private topic. The method for determining user types is described below and will not be repeated here.
[0060] S302 determines multiple candidate topics to be pushed based on public domain traffic.
[0061] By acquiring public domain traffic and performing data analysis on it, topics of public interest can be extracted. One possible implementation involves acquiring data related to the information being pushed from public domain traffic, and then extracting candidate topics from this data. For example, if an automaker needs to push information, it can acquire vehicle-related data from public domain traffic and then extract vehicle-related topics of public interest. For instance, it can crawl text information about popular online topics from various social media platforms, short video platforms, and specific vehicle-related vertical websites (such as influential automotive vertical websites / forums like Autohome, Xcar, Sina Auto, and Sohu Auto). The crawled data can then be cleaned, and keywords can be extracted to obtain multiple public domain topics.
[0062] After obtaining public domain topics, the popularity of each topic can be determined. The popularity of a public domain topic can be determined based on its frequency of appearance in public domain traffic. These topics can be sorted according to their popularity, and then the top-ranked topics can be selected as candidate topics. For example, a public social platform may include a trending topics list, and the ranking of public domain topics on this list can be used to represent their popularity. The top 10 trending topics can be used as candidate topics.
[0063] S303, based on one or more of the private domain topics, determine a target topic from a plurality of candidate topics, wherein the target topic has a corresponding target private domain topic, and the target private domain topic is a private domain topic among one or more private domain topics whose similarity to the target topic is greater than a preset threshold.
[0064] When pushing information, the similarity between each candidate topic and each private domain topic can be calculated, and then the target topic can be determined based on the similarity. For example, a similarity greater than a preset threshold can be used as the target similarity. The candidate topic and private domain topic corresponding to the target similarity are determined, and then the candidate topic is used as the target topic, and the private domain topic is used as the target private domain topic.
[0065] In one possible implementation, the target topic can include one. For example, alternative topics and private domain topics corresponding to the maximum similarity can be determined, and then the alternative topic can be used as the target topic, and the private domain topic can be used as the target private domain topic.
[0066] In one possible implementation, there can be multiple target topics, each of which can have a corresponding target private domain topic. For each target topic, information can be pushed using steps S304-S305.
[0067] When calculating the similarity between each candidate topic and each private topic, natural language processing techniques, such as word vectors, text classification, and clustering algorithms, can be used to compare the similarity between the two topics; or graph theory algorithms, such as node degree, path length, and shortest path, can be used to compare the relevance between the two topics.
[0068] S304, Generate a list of users corresponding to the target topic based on the feature value of the target private domain topic for each user.
[0069] Based on the characteristic values of each user's target private domain topic, the user type under the target topic can be determined. The user types include key seed users, potential users, extreme users, and non-reachable users.
[0070] Figure 4 This is a schematic diagram illustrating a user classification method provided in an embodiment of this application. For example... Figure 4 As shown, if the first feature value of the target private domain topic is greater than the first threshold, and the second feature value of the target private domain topic is greater than the second threshold, then the user is interested in the target topic and has a positive attitude, and the user type can be identified as a key seed user; if the first feature value of the target private domain topic is greater than the first threshold, and the second feature value of the target private domain topic is less than or equal to the second threshold, then the user is interested in the target topic but has a negative attitude, and the user type can be identified as an extreme user; if the first feature value of the target private domain topic is less than or equal to the first threshold, and the second feature value of the target private domain topic is greater than the second threshold, then the user is not interested in the target topic but has a positive attitude, and the user type can be identified as a potential user; if the first feature value of the target private domain topic is less than or equal to the first threshold, and the second feature value of the target private domain topic is less than or equal to the second threshold, then the user is not interested in the target topic and has a negative attitude, and the user type can be identified as an unreached user.
[0071] Based on the user type corresponding to each user under the target topic, a user list can be generated. All key seed users of the target topic form a key seed user list; all potential users of the target topic form a potential user list; all extreme users of the target topic form an extreme user list; and all unreachable users of the target topic form an unreachable user list.
[0072] S305, Based on the user list, push the target topic.
[0073] A target topic can include multiple different push notifications, each with a different purpose. For example, some push notifications might be aimed at key seed users to promote the product, others at developing potential users, and still others at conducting user research to make technological improvements. Operations personnel can plan different push notifications for different user lists based on the target topic, ensuring that each push notification has a distinct purpose. Based on these different purposes, the push notifications can then be sent to different users.
[0074] When sending push notifications, multiple notifications can be retrieved for a specific target topic, each with a specific purpose. These notifications can be generated automatically or entered by operations personnel.
[0075] Based on the purpose of the push notification, a target push list is determined; the push notification is then sent to users in the target push list. For example, the push notification may include a first push notification corresponding to key seed users and a second push notification corresponding to potential users; when sending the notification, the first push notification can be sent to users in the key seed user list, and the second push notification can be sent to users in the potential user list.
[0076] In this application embodiment, users can be segmented based on their interests and emotional inclinations to obtain different user lists, and different push information can be pushed to different user lists to avoid pushing marketing information to biased or indifferent users. This can achieve personalized and highly participatory activity promotion to meet the activity needs of different groups.
[0077] When operating and promoting based on the methods in this application embodiment, each user can be categorized into different user lists based on their preferences for various private domain topics. To improve information reach during push notifications, different push notifications can be generated for different topics, thus delivering information to each user according to the topic and the corresponding push notification purpose. Users can receive push notifications on topics they are interested in, achieving a better promotional effect. When technical improvements are needed, corresponding Q&A survey reports can be pushed to users with negative attitudes to obtain effective user feedback. Improvements based on this feedback can make the improved product more appealing to more users. Based on the methods in this application embodiment, different activity topics can be planned for different users during information push notifications, thereby meeting the information needs of different groups, improving the reach of push notifications, and making marketing activities more precise and effective.
[0078] It should be noted that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.
[0079] Reference Figure 5 The diagram illustrates an information push device according to an embodiment of this application, which may specifically include a tag information determination module 51, a candidate topic determination module 52, a target topic determination module 53, a user list generation module 54, and a target topic push module 55, wherein:
[0080] The tag information determination module 51 is used to determine the tag information of each user based on user behavior data in the private domain traffic. The tag information includes one or more feature values corresponding to the private domain topics. The feature values are used to characterize the user's tendency towards the corresponding private domain topics.
[0081] The alternative topic determination module 52 is used to determine multiple alternative topics for push based on public domain traffic;
[0082] The target topic determination module 53 is used to determine a target topic from a plurality of candidate topics based on one or more of the private domain topics. The target topic corresponds to one of the private domain topics, and the target private domain topic is a private domain topic among one or more private domain topics whose similarity to the target topic is greater than a preset threshold.
[0083] User list generation module 54 is used to generate a user list corresponding to the target topic based on the feature value of each private domain topic.
[0084] The target topic push module 55 is used to push the target topic based on the user list.
[0085] In one possible implementation, the label information determination module 51 includes:
[0086] The data acquisition submodule is used to acquire the user's interaction behavior data and access frequency data for each private domain topic from the user behavior data;
[0087] The first feature value calculation submodule is used to calculate the first feature value of each private domain topic based on the interaction behavior data and the access frequency data of each private domain topic. The first feature value is used to characterize the user's interest tendency towards the corresponding private domain topic.
[0088] The second feature value calculation submodule is used to perform sentiment analysis on the user behavior data using a sentiment analysis algorithm, and to determine the second feature value for each private domain topic. The second feature value is used to characterize the user's sentiment tendency towards the corresponding private domain topic.
[0089] The feature value determination submodule is used to determine the feature value of each of the private domain topics based on the first feature value and the second feature value.
[0090] In one possible implementation, the user list generation module 54 includes:
[0091] The user type determination submodule is used to determine the user type of the user under the target topic based on the feature value of each private domain topic. The user type includes key seed users, potential users, extreme users, and non-reachable users.
[0092] The user list generation submodule is used to generate the user list according to the user type. The user list includes a key seed user list, a potential user list, a biased user list, and / or an unreachable user list.
[0093] In one possible implementation, the aforementioned user type determination submodule includes:
[0094] A determining unit is configured to determine a target private domain topic that matches the target topic from one or more of the private domain topics;
[0095] The key seed user determination unit is used to determine the user type as a key seed user if the first feature value of the target private domain topic is greater than a first threshold and the second feature value of the target private domain topic is greater than a second threshold.
[0096] The radical user determination unit is used to determine the user type as a radical user if the first feature value of the target private domain topic is greater than the first threshold and the second feature value of the target private domain topic is less than or equal to the second threshold.
[0097] A potential user determination unit is configured to determine the user type as a potential user if the first feature value of the target private domain topic is less than or equal to the first threshold and the second feature value of the target private domain topic is greater than the second threshold.
[0098] The non-reachable user determination unit is used to determine the user type as a non-reachable user if the first feature value of the target private domain topic is less than or equal to the first threshold and the second feature value of the target private domain topic is less than or equal to the second threshold.
[0099] In one possible implementation, the above-mentioned alternative topic determination module 52 includes:
[0100] The public domain topic extraction submodule is used to extract multiple public domain topics from the public domain traffic and determine the topic popularity of each public domain topic.
[0101] The alternative topic filtering submodule is used to filter out multiple alternative topics from multiple public domain topics based on the topic popularity.
[0102] In one possible implementation, the target topic determination module 53 includes:
[0103] A similarity calculation submodule is used to calculate the similarity between each of the candidate topics and the private domain topics;
[0104] The target topic determination submodule is used to select candidate topics with a similarity greater than a preset threshold as the target topic.
[0105] In one possible implementation, the target topic push module 55 includes:
[0106] The push information acquisition submodule is used to acquire multiple push information messages corresponding to the target topic, and each push information message has a corresponding push purpose;
[0107] The target push list determination submodule is used to determine the target push list corresponding to the push information based on the push purpose;
[0108] The push module is used to push the push information to users in the target push list.
[0109] As the apparatus embodiments are basically similar to the method embodiments, they are described in a relatively simple manner. For relevant details, please refer to the description in the method embodiment section.
[0110] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 6 As shown, the electronic device 600 of this embodiment includes: at least one processor 60 ( Figure 6 (Only one is shown in the diagram), memory 61, and computer program 62 stored in said memory 61 and executable on said at least one processor 60, which, when executed, implements the steps in any of the above method embodiments.
[0111] The electronic device 600 may be a server, a cloud server, etc. This electronic device may include, but is not limited to, a processor 60 and a memory 61. Those skilled in the art will understand that... Figure 6 This is merely an example of electronic device 600 and does not constitute a limitation on electronic device 600. It may include more or fewer components than shown, or combine certain components, or different components, such as input / output devices, network access devices, etc.
[0112] The processor 60 may be a Central Processing Unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.
[0113] In some embodiments, the memory 61 may be an internal storage unit of the electronic device 600, such as a hard disk or memory of the electronic device 600. In other embodiments, the memory 61 may be an external storage device of the electronic device 600, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the electronic device 600. Furthermore, the memory 61 may include both internal and external storage units of the electronic device 600. The memory 61 is used to store the operating system, applications, bootloader, data, and other programs, such as the program code of the computer program. The memory 61 can also be used to temporarily store data that has been output or will be output.
[0114] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps described in the various method embodiments above.
[0115] This application provides a computer program product that, when run on an electronic device, enables the electronic device to perform the steps described in the various method embodiments above.
[0116] The embodiments described above are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.
Claims
1. An information push method, characterized in that, include: Based on user behavior data in private domain traffic, the tag information of each user is determined. The tag information includes one or more feature values corresponding to private domain topics. The feature values are used to characterize the user's tendency towards the corresponding private domain topic. Based on public domain traffic, several alternative topics to be pushed are determined; Based on one or more of the private domain topics, a target topic is determined from a plurality of candidate topics. The target topic has a corresponding target private domain topic. The target private domain topic is a private domain topic among one or more private domain topics whose similarity to the target topic is greater than a preset threshold. Based on the feature value of the target private domain topic for each user, a user list corresponding to the target topic is generated; Based on the user list, the target topics are pushed out; The step of generating a user list corresponding to the target private domain topic based on the feature value of each user's target private domain topic includes: Based on the feature values of the target private domain topic for each user, determine the user type of each user under the target topic; Based on the user type, the user list is generated, which includes a key seed user list, a potential user list, a biased user list, and / or an unreachable user list. Specifically, pushing the target topic based on the user list includes: Obtain multiple push notifications corresponding to the target topic, each of which has a corresponding push purpose; Based on the stated push purpose, determine the target push list corresponding to the push information; The push notification is sent to the users in the target push list.
2. The method as described in claim 1, characterized in that, The process of determining the tag information for each user based on user behavior data in private domain traffic includes: The user's interaction behavior data and access frequency data for each private domain topic are obtained from the user behavior data. Based on the interaction behavior data and access frequency data of each private domain topic, a first feature value is calculated for each private domain topic. The first feature value is used to characterize the user's interest tendency towards the corresponding private domain topic. Sentiment analysis is performed on the user behavior data to determine a second feature value for each private domain topic. The second feature value is used to characterize the user's sentiment towards the corresponding private domain topic. The feature value of each private domain topic is determined based on the first feature value and the second feature value.
3. The method as described in claim 1, characterized in that, The step of determining the user type of each user under the target private domain topic based on the feature value of each user includes: If the first feature value of the target private domain topic is greater than the first threshold, and the second feature value of the target private domain topic is greater than the second threshold, then the user type is determined to be a key seed user. If the first feature value of the target private domain topic is greater than the first threshold, and the second feature value of the target private domain topic is less than or equal to the second threshold, then the user type is determined to be an extreme user. If the first feature value of the target private domain topic is less than or equal to the first threshold, and the second feature value of the target private domain topic is greater than the second threshold, then the user type is determined to be a potential user. If the first feature value of the target private domain topic is less than or equal to the first threshold, and the second feature value of the target private domain topic is less than or equal to the second threshold, then the user type is determined to be an unreachable user.
4. The method according to any one of claims 1-3, characterized in that, The method of determining multiple candidate topics for information push based on public domain traffic includes: Extract multiple public domain topics from the public domain traffic and determine the popularity of each public domain topic; Based on the popularity of the topic, multiple candidate topics are selected from the various public domain topics.
5. The method according to any one of claims 1-3, characterized in that, The step of determining the target topic from a plurality of candidate topics based on one or more of the private domain topics includes: Calculate the similarity between each of the candidate topics and the private domain topics; Candidate topics with a similarity greater than a preset threshold are selected as the target topics.
6. An information push device, characterized in that, include: The tag information determination module is used to determine the tag information of each user based on user behavior data in private domain traffic. The tag information includes one or more private domain topics and feature values of each private domain topic. The feature values are used to characterize the user's tendency towards the corresponding private domain topic. The alternative topic determination module is used to determine multiple alternative topics to be pushed based on public domain traffic; The target topic determination module is used to determine a target topic from multiple candidate topics based on one or more private domain topics. The target topic has a corresponding target private domain topic, and the target private domain topic is a private domain topic among one or more private domain topics whose similarity to the target topic is greater than a preset threshold. The user list generation module is used to generate a user list corresponding to the target topic based on the feature value of the target private domain topic for each user. The target topic push module is used to push the target topic based on the user list; The user list generation module is used for: Based on the feature values of the target private domain topic for each user, determine the user type of each user under the target topic; Based on the user type, the user list is generated, which includes a key seed user list, a potential user list, a biased user list, and / or an unreachable user list. The target topic push module is used for: Obtain multiple push notifications corresponding to the target topic, each of which has a corresponding push purpose; Based on the stated push purpose, determine the target push list corresponding to the push information; The push notification is sent to the users in the target push list.
7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method as described in any one of claims 1-5.
8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-5.