A media fusion information pushing system

Abstract

This invention relates to the field of data processing technology, and more particularly to a converged media information push system, comprising a converged media information module, a traffic statistics module, a user tree module, a full push module, and a personalized push module. The converged media information module is responsible for integrating medical news from different platforms, categorizing it by disease type, and tagging and pushing information based on traffic data. The full push module pushes information broadly or comprehensively to corresponding branches of the user tree based on the tag type. The personalized push module provides customized content for specific user branches, recommending content based on user behavior. The traffic statistics module monitors user interaction in real time, while the user tree module dynamically adjusts the user's position in the user tree based on their browsing behavior. The system can accurately pinpoint user preferences and provide medical news content. Through these interconnected modules, the system improves its ability to process large-scale data while enhancing the real-time nature and comprehensiveness of recommendations.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and more particularly to a converged media information push system. Background Technology

[0002] With the rapid development of internet technology, converged media has become an important means of information dissemination. Converged media effectively integrates various media formats such as text, images, video, and audio, providing richer and more interactive ways to deliver information. Especially in the healthcare field, the application of converged media can greatly improve the public's understanding and acceptance of health knowledge. However, in this process, how to effectively push relevant medical news and health information to target users has become a key challenge.

[0003] Currently, many information push systems rely on collaborative filtering algorithms to achieve personalized content delivery. Collaborative filtering algorithms make recommendations by analyzing the similarity between users and their content preferences. However, this method has some significant limitations in converged media information push, including the cold start problem: for new users or new content, collaborative filtering algorithms often struggle to provide effective recommendations immediately due to a lack of sufficient user behavior data for analysis. The sparsity problem: with large-scale users and content, the user-content interaction matrix is ​​often very sparse, making it difficult to find similar users or related content. Collaborative filtering algorithms often face performance bottlenecks when processing large numbers of users and content. As the number of users and the amount of content increase, the cost of calculating similarity between users or between users and content rises sharply, leading to a decrease in algorithm efficiency. In a converged media environment, content is updated frequently and diverse, and user behavior and preferences are constantly changing. Collaborative filtering algorithms typically require a long time to process these dynamic changes, making it difficult to achieve real-time personalized recommendations. Over-specialization: Collaborative filtering algorithms tend to recommend content highly correlated with users' past behavior, potentially leading to overly concentrated recommendations in terms of topic or type, lacking comprehensiveness. To address these performance shortcomings, the converged media information push system proposed in this invention adopts an innovative combination of architecture and algorithms, aiming to improve the ability to process large-scale data while enhancing the real-time and comprehensiveness of recommendations. Summary of the Invention

[0004] To address the above problems, this invention provides a converged media information push system.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] A converged media information push system includes a converged media information module, a traffic statistics module, a user tree module, a full push module, and a personalized push module. The converged media information module is connected to the full push module and the personalized push module, the user tree module is connected to the full push module and the personalized push module, and the traffic statistics module is connected to the converged media information module and the user tree module.

[0007] The converged media information module is used to store medical news information from several platforms, classify the medical news information and its associated traffic information according to disease type, and send the medical news information to the full push module and the personalized push module based on the traffic information.

[0008] The full push module is used to globally push medical news information to all branches in the user tree module.

[0009] The personalized push module is used to push information to branch users in the user tree module.

[0010] The traffic statistics module is used to monitor traffic behavior in the user tree module, generate traffic data associated with medical news information based on traffic behavior, and synchronize it to the converged media information module.

[0011] The user tree module is used to generate a user tree based on disease type and adjust the branch affiliation of users based on traffic behavior.

[0012] Furthermore, the step of sending medical news information to the full push module and the personalized push module based on traffic information includes:

[0013] If the medical news information is not associated with traffic data, then the medical news information will be marked as a trial delivery and sent to the full push module;

[0014] If medical news information is associated with traffic data, and the prevalence of the traffic data is higher than the preset value, then the medical news information will be marked as being sent to the full push module.

[0015] If medical news information is associated with traffic data, and the prevalence of the traffic data is lower than a preset value, then the medical news information will be marked with several branches based on the associated traffic data and sent to the personalized push module.

[0016] Furthermore, the global push of medical news information to all branches in the user tree module includes:

[0017] The full push module distributes medical news information marked as trial delivery to a preset number of users across all branches of the user tree;

[0018] The full push module delivers medical news information marked as "full delivery" to all users across all branches of the user tree.

[0019] Furthermore, the traffic behavior in the monitored user tree module includes:

[0020] Real-time monitoring of user click behavior, completion rate, and forwarding behavior.

[0021] Furthermore, the traffic data associated with generating medical news information based on traffic behavior includes:

[0022] The click-through rate, completion rate, and forwarding rate of users are statistically analyzed to obtain the branch click-through rate, completion rate, and forwarding rate of the user tree, and then correlated with medical news information.

[0023] Furthermore, the step of performing several branch markers includes:

[0024] The recommendation score is obtained by weighting the click-through rate, completion rate, and forwarding rate of the user tree branches.

[0025] If the recommendation score is greater than or equal to the preset recommendation value, the branch will be marked as recommended.

[0026] If the recommended score is less than the preset recommended value, it will not be marked.

[0027] Furthermore, the branch push to branch users in the user tree module includes branch push of medical news information based on recommendation tags.

[0028] Furthermore, adjusting user branch affiliation based on traffic behavior includes:

[0029] Create user profiles for users;

[0030] Based on the initial browsing behavior of new users, users are categorized into the corresponding disease category branches;

[0031] Continuously monitor users' browsing behavior under the categorized disease types and adjust users' branch affiliation accordingly.

[0032] Furthermore, the disease category branch includes several disease sub-branches.

[0033] The beneficial effects of this invention are as follows: This invention stores and categorizes various medical news information through a converged media information module, enabling the system to effectively push new content based on its category and attributes. Furthermore, by analyzing new users' initial reactions to different categories of content, the system can quickly build a preliminary understanding of user preferences, thereby alleviating the cold start problem for new users. The traffic statistics module monitors real-time user interactions with content, such as clicks, views, and forwarding behaviors. Through the analysis of this real-time data, the system can more accurately capture users' current interests, enabling effective recommendations even with limited user-content interaction data. The user tree module dynamically generates and adjusts user trees based on user behavior, effectively managing large amounts of user data. Through the user tree model, the system can efficiently process user grouping, ensuring recommendation accuracy on a large user base. The full-push module pushes information to a broad range of users, while the personalized push module provides customized content based on specific user behavior data. This combination not only ensures timely updates of recommended content but also guarantees the diversity and personalization of content recommendations. By segmenting users using a user tree, the computational calculation of user similarity is reduced, thus saving computing resources. Furthermore, the separate execution of the full push module and the personalized push module reduces logical judgments and further optimizes performance. This effectively improves the ability to process large-scale data, providing users with a more efficient, accurate, and comprehensive medical and health information push service. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the structure of a converged media information push system according to the present invention.

[0035] Figure 2 This is a flowchart illustrating the steps of adjusting a user's branch affiliation based on traffic behavior in this invention. Detailed Implementation

[0036] Please see Figure 1-2 As shown, this invention relates to a converged media information push system;

[0037] Specifically, the present invention provides a converged media information push system, including a converged media information module, a traffic statistics module, a user tree module, a full push module, and a personalized push module. The converged media information module is connected to the full push module and the personalized push module, the user tree module is connected to the full push module and the personalized push module, and the traffic statistics module is connected to the converged media information module and the user tree module.

[0038] The converged media information module is used to store medical news information from several platforms, classify the medical news information and its associated traffic information according to disease type, and send the medical news information to the full push module and the personalized push module based on the traffic information.

[0039] The full push module is used to globally push medical news information to all branches in the user tree module.

[0040] The personalized push module is used to push information to branch users in the user tree module.

[0041] The traffic statistics module is used to monitor traffic behavior in the user tree module, generate traffic data associated with medical news information based on traffic behavior, and synchronize it to the converged media information module.

[0042] The user tree module is used to generate a user tree based on disease type and adjust the branch affiliation of users based on traffic behavior.

[0043] It's important to note that the converged media information module, at the software level, is responsible for collecting, storing, and categorizing medical news content from different platforms. It uses advanced database technologies (such as MySQL or MongoDB) to store large amounts of text, images, video, and audio information. This information is categorized by disease type; for example, news and research related to heart disease are grouped into one category. This module also utilizes natural language processing techniques (such as NLP) to analyze text content and extract keywords to aid in information categorization and retrieval. The traffic statistics module primarily monitors and analyzes user interactions with content within the converged media information module. It uses data analysis and machine learning algorithms (e.g., Pandas and Scikit-learn libraries in Python) to process and parse user click behavior, viewing time, and sharing behavior data. This data is then used to generate traffic reports, helping the system understand which content is more popular with users. The user tree module, at the software level, uses graph-based data structure algorithms to generate user trees representing different user groups and their preferences. It dynamically adjusts the user's position and category in the tree based on user interaction data (obtained from the traffic statistics module) to ensure the accuracy of personalized recommendations. The full push module is responsible for widely disseminating medical news information to all users. At the hardware level, it can rely on high-performance servers and network infrastructure to support large-scale data transmission and processing. At the software level, advanced programming techniques (such as multithreading and asynchronous processing) can be used to optimize the efficiency of the push process.

[0044] Furthermore, the step of sending medical news information to the full push module and the personalized push module based on traffic information includes:

[0045] If the medical news information is not associated with traffic data, then the medical news information will be marked as a trial delivery and sent to the full push module;

[0046] If medical news information is associated with traffic data, and the prevalence of the traffic data is higher than the preset value, then the medical news information will be marked as being sent to the full push module.

[0047] If medical news information is associated with traffic data, and the prevalence of the traffic data is lower than a preset value, then the medical news information will be marked with several branches based on the associated traffic data and sent to the personalized push module.

[0048] It's important to note that for medical news information not yet linked to traffic data, the system marks them as trial delivery. This means the information is new or hasn't yet been viewed and interacted with by a sufficient number of users. In this case, the news is sent to the full push module to reach a broad user base. For example, if a report about a new drug has just been published, the system will classify it as trial delivery to assess its popularity among a wider user group. When medical news information is already linked to traffic data, and this data shows a news prevalence rate higher than a preset threshold, the system marks it as full delivery. In this case, the news is considered broadly appealing and suitable for push to all users. For example, a news story about a major public health event, if it shows high click-through rates and widespread user attention in the initial trial delivery, the system will upgrade it to full delivery to ensure that every user receives this important information. For medical news information that is linked to traffic data but has a prevalence rate lower than the preset value, the system will branch-mark this information and send it to the personalized push module. This indicates that while this information may not have attracted widespread attention from a broad user base, it may still be highly relevant to a specific user group. For example, a research report on a specific rare disease may only be of high value to patients or specialists with that disease, so the system will push this type of information to the relevant user branches.

[0049] Furthermore, the global push of medical news information to all branches in the user tree module includes:

[0050] The full push module distributes medical news information marked as trial delivery to a preset number of users across all branches of the user tree;

[0051] The full push module delivers medical news information marked as "full delivery" to all users across all branches of the user tree.

[0052] It's important to note that when medical news is marked as a trial release, the full-push module is responsible for delivering this information to a predetermined number of users in each branch of the user tree. This predetermined target audience strategy aims to assess the acceptance and impact of the news information among different user groups. For example, suppose a news item about a novel flu vaccine has just been released. Due to a lack of sufficient historical traffic data, the system classifies it as a trial release. The full-push module will push this news to a certain percentage of users in each branch of the user tree, for example, 30% of users in each branch. This allows for the collection of initial user reactions and traffic data regarding the news without excessively impacting the user experience.

[0053] When medical news is marked as "full delivery," it indicates that the information is highly relevant and valuable to a broad user base. In this case, the full delivery module pushes this news to all users at every branch of the user tree. For example, an update on a major public health guideline might be marked as "full delivery" because it is extremely important to all users. The full delivery module then ensures that this information is pushed to all users at every branch of the user tree to ensure that everyone receives this critical health information in a timely manner.

[0054] Furthermore, the global push of medical news information to all branches in the user tree module includes:

[0055] The full push module distributes medical news information marked as trial delivery to a preset number of users across all branches of the user tree;

[0056] The full push module delivers medical news information marked as "full delivery" to all users across all branches of the user tree.

[0057] When medical news information is marked as a trial release, it usually means that this information is newly released and has not yet accumulated enough user interaction and traffic data to assess its general audience and interest. In this case, the task of the full push module is to distribute these trial news messages to a predetermined number of users in each branch of the user tree module. First, the full push module will identify all branches in the user tree module. Each branch represents a group of users with a preference for information about a specific disease. For each branch, the system will calculate a predetermined target number of users, which is usually based on a proportion of the total number of users in that branch. For example, the system may decide to push the trial news information to 20% of users in each branch. Then, users who meet the predetermined proportion in each branch will be randomly selected, and the trial medical news information will be sent to these selected users. This random selection method helps ensure that the impact assessment of the trial release is representative and fair. In this way, the full push module can effectively conduct preliminary market testing and user response collection for the trial medical news information without affecting the overall system performance and user experience. This provides important data support for subsequent full push and personalized push.

[0058] News delivered to the entire user tree is sent to every branch, regardless of their specific interests or preferences. This ensures that everyone, from users highly concerned about their health to those who only occasionally check health information, receives this critical information. Unlike trial delivery strategies, which involve random selection and pre-selected recipients, the full delivery strategy does not perform any screening but directly pushes information to all members of every branch of the user tree. This means that every user, regardless of their previous interaction frequency or interests, will receive medical news deemed highly important. Because full delivery information often involves significant or urgent health information, the system may assign it higher priority. This means that this information may be placed in a more prominent position in a user's news feed queue to ensure their attention. For example, if there is news about a widespread infectious disease outbreak or an important public health update, the system will mark this information as full delivery and ensure that every user receives the message, regardless of which branch of the user tree they are on. This strategy ensures that critical information is disseminated rapidly and widely in emergencies or situations involving the health and safety of a large number of users.

[0059] Furthermore, the traffic behavior in the monitored user tree module includes:

[0060] Real-time monitoring of user click behavior, completion rate, and forwarding behavior.

[0061] In some embodiments, the traffic statistics module tracks user click behavior on medical news content in real time. This includes recording the number of times and the timing of clicks on news links. For example, if a user clicks on a news article about heart disease research while browsing a news list, the traffic statistics module will record this action. This monitoring helps understand the level of user interest in different medical news topics. For medical news containing video or audio content, the traffic statistics module will monitor user completion behavior, i.e., whether the user watches or listens to the entire content. This metric is important for assessing user engagement and depth of interest in specific content. For example, whether a user watches the entire video report on new drug development can provide the system with in-depth insights into user interests. The module also tracks user sharing behavior of medical news, including the number of times and patterns of users sharing news through social media, email, or other means. This behavior often indicates that users are not only interested in the content but also believe that the information is valuable to others. For example, a user might share a news article about healthy lifestyles with family or friends.

[0062] Further, the branching markers include:

[0063] The recommendation score is obtained by weighting the click-through rate, completion rate, and forwarding rate of the user tree branches.

[0064] If the recommendation score is greater than or equal to the preset recommendation value, the branch will be marked as recommended.

[0065] If the recommended score is less than the preset recommended value, it will not be marked.

[0066] Furthermore, the branch push to branch users in the user tree module includes branch push of medical news information based on recommendation tags.

[0067] In some embodiments, a weighted average score, or recommendation score, is first calculated based on the click-through rate, completion rate, and sharing rate of each branch of the user tree. These metrics reflect users' interest in and engagement with specific content. For example, a high click-through rate indicates user interest in this type of content, a high completion rate indicates users are willing to spend time watching or listening to the content in its entirety, and a high sharing rate indicates users believe the content is worth sharing. By combining these different metrics, the system can derive a comprehensive evaluation score that reflects the overall response and preference of each branch for a certain type of content. The calculated recommendation score is used to determine whether a branch should be labeled as a recommendation. If a branch's recommendation score reaches or exceeds a preset recommendation threshold, the system will label that branch as a recommendation. This means that users on that branch have shown high interest and engagement with this type of content, and therefore this type of content is suitable to be pushed to users on that branch. If a branch's recommendation score is below the preset recommendation threshold, the system will not label that branch. This means that users on that branch have insufficient interest or engagement with this type of content, and therefore this type of content is not suitable to be pushed to users on that branch.

[0068] Furthermore, adjusting user branch affiliation based on traffic behavior includes:

[0069] Create user profiles for users;

[0070] Based on the initial browsing behavior of new users, users are categorized into the corresponding disease category branches;

[0071] Continuously monitor users' browsing behavior under the categorized disease types and adjust users' branch affiliation accordingly.

[0072] Furthermore, the disease category branch includes several disease sub-branches.

[0073] In some embodiments, a detailed user profile is first created for each newly registered or joined user. This profile contains the user's basic information, preferences, and initial browsing behavior. The user profile forms the basis of the personalized recommendation system, helping the system understand each user's unique needs and interests. Next, the new user is categorized into the corresponding disease category branch based on their initial browsing behavior. For example, if a new user frequently browses articles and videos related to heart disease after joining the system, the system will categorize this user into the heart disease-related branch. This initial categorization, based on the user's initial exhibited interests and behavioral patterns, provides direction for subsequent personalized recommendations. User browsing behavior within their respective disease category branch is tracked and analyzed in real time. This includes clicks, reading time, completion rates, and interactions (such as likes, comments, and shares) on specific medical news and content. For example, if a user was initially categorized in the heart disease branch but subsequently begins frequently reading information about diabetes, this change will be captured. The system uses data analysis tools and algorithms (such as behavioral pattern recognition algorithms) to identify trends in user behavior. These tools can process and interpret large amounts of user behavior data to accurately understand the evolution of user interests. For example, the system might detect a user gradually reducing their browsing of heart disease content while increasing their interest in diabetes-related content. Based on the analysis of these changes in user behavior patterns, the system dynamically adjusts the user's branch affiliation. If it detects a shift in a user's interest from one disease category to another, the user will be moved from one branch to the other accordingly. In the previous example, this means moving the user from the heart disease branch to the diabetes branch. After this adjustment, the system provides more accurate and personalized medical news recommendations based on the user's new branch affiliation. This ensures that the recommended content not only reflects the user's latest interests but also better meets their current information needs. To further improve the accuracy and relevance of the recommendations, disease category branches are subdivided into several sub-branches. These sub-branches represent more specific disease categories or related health topics. For example, under the broad category of heart disease, there might be sub-branches such as hypertension and coronary heart disease, each providing more precise content recommendations for a specific disease or health issue.

[0074] The above embodiments are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A converged media information push system, characterized in that, It includes a converged media information module, a traffic statistics module, a user tree module, a full push module, and a personalized push module. The converged media information module is connected to the full push module and the personalized push module, the user tree module is connected to the full push module and the personalized push module, and the traffic statistics module is connected to the converged media information module and the user tree module. The converged media information module is used to store medical news information from several platforms, classify the medical news information and its associated traffic information according to disease type, and send the medical news information to the full push module and the personalized push module based on the traffic information. The full push module is used to globally push medical news information to all branches in the user tree module. The personalized push module is used to push information to branch users in the user tree module. The traffic statistics module is used to monitor traffic behavior in the user tree module, generate traffic data associated with medical news information based on traffic behavior, and synchronize it to the converged media information module. The user tree module is used to generate a user tree based on disease type and adjust the branch affiliation of users based on traffic behavior; The method of sending medical news information to the full push module and personalized push module based on traffic information includes: If the medical news information is not associated with traffic data, then the medical news information will be marked as a trial delivery and sent to the full push module; If medical news information is associated with traffic data, and the prevalence of the traffic data is higher than the preset value, then the medical news information will be marked as being sent to the full push module. If medical news information is associated with traffic data, and the prevalence of the traffic data is lower than a preset value, then the medical news information will be marked with several branches based on the associated traffic data and sent to the personalized push module. The traffic behavior monitored in the user tree module includes: Real-time monitoring of user click behavior, completion rate, and forwarding behavior; The step of marking several branches includes: The recommendation score is obtained by weighting the click-through rate, completion rate, and forwarding rate of the user tree branches. If the recommendation score is greater than or equal to the preset recommendation value, the branch will be marked as recommended. If the recommended score is less than the preset recommended value, it will not be marked.

2. The converged media information push system according to claim 1, characterized in that, The global push of medical news information to all branches in the user tree module includes: The full push module distributes medical news information marked as trial delivery to a preset number of users across all branches of the user tree; The full push module delivers medical news information marked as "full delivery" to all users across all branches of the user tree.

3. The converged media information push system according to claim 1, characterized in that, The traffic data associated with generating medical news information based on traffic behavior includes: The click-through rate, completion rate, and forwarding rate of users are statistically analyzed to obtain the branch click-through rate, completion rate, and forwarding rate of the user tree, and then correlated with medical news information.

4. The converged media information push system according to claim 1, characterized in that, The branch push to branch users in the user tree module includes branch push of medical news information based on recommendation tags.

5. A converged media information push system according to claim 1, characterized in that, The method of adjusting a user's branch affiliation based on traffic behavior includes: Create user profiles for users; Based on the initial browsing behavior of new users, users are categorized into the corresponding disease category branches; Continuously monitor users' browsing behavior under the categorized disease types and adjust users' branch affiliation accordingly.

6. A converged media information push system according to claim 5, characterized in that, The disease category branch includes several disease sub-branches.

Images