An agent-based 5G message personalized push method

Abstract

This invention discloses a personalized 5G message push method based on intelligent agents, belonging to the field of message push technology. The push steps are as follows: S1, establish a mobile phone information database, obtain different mobile phone information versions and hardware configuration data, and intelligently classify the mobile phones into smartphones and feature phones; S2, obtain real-time reading feedback of different push messages, analyze and judge, and obtain user preferred message types; S3, generate personalized push messages for different users, intelligently push 5G messages according to different mobile phone types and different user preferences, and push concise text messages to identified feature phones. This invention optimizes the user experience for the elderly, and through continuous data collection and strategy optimization, it quickly adapts to changes in market and user needs, enabling the platform to maintain a leading position in the competition, provide users with high-quality services, accurately match users' current interests and needs, and increase user engagement.

Description

Technical Field

[0001] This invention relates to the field of message push technology, specifically to a 5G message personalized push method based on intelligent agents. Background Technology

[0002] With the rapid development and widespread application of 5G technology, 5G messaging, with its rich media and strong interactivity, has become an important channel for information dissemination and service delivery. Intelligent agents are entities with intelligent characteristics that can perceive the environment and make decisions and take actions based on their own goals. The integration of intelligent agents into 5G messaging is usually based on artificial intelligence technology, such as large models. For example, China Unicom has built 5G messaging intelligent agents based on the Yuanjing Large Model MaaS platform. These intelligent agents can understand users' natural language commands and provide services such as intelligent question answering and business processing, giving 5G messaging more powerful intelligent interaction capabilities.

[0003] In practical applications, users use a variety of mobile devices with different performance levels, and their preferences for message content also vary greatly. How to achieve accurate and efficient personalized 5G message push has become an urgent problem to be solved.

[0004] Currently, most existing push notification technologies lack detailed differentiation based on the characteristics of mobile devices. On the one hand, they fail to fully consider the significant differences between smartphones and feature phones in terms of hardware configuration, user interface, and display capabilities. Feature phones generally have small screens, low resolution, and simple functions. If messages with the same format and content as smartphones are directly pushed to them, display abnormalities and reading difficulties are likely to occur, seriously affecting the user experience. On the other hand, traditional push notification technologies do not analyze user preferences in depth or in real time, often relying only on limited historical data or general tags, making it difficult to accurately match users' current interests and needs, resulting in low message open rates and insufficient user engagement. To address this, we propose a personalized 5G message push method based on intelligent agents. Summary of the Invention

[0005] To address the aforementioned technical problems, a personalized 5G message push method based on intelligent agents is provided. This technical solution solves the problems mentioned above.

[0006] To achieve the above objectives, the technical solution adopted by this invention is: a 5G message personalized push method based on intelligent agents, wherein the push steps are as follows:

[0007] S1. Establish a mobile phone information database, obtain information on different mobile phone versions and hardware configuration data, and perform intelligent classification to divide mobile phones into smartphones and feature phones for the elderly.

[0008] S2. Obtain real-time reading feedback on different push messages, analyze and judge the user's preferred message types;

[0009] S3 generates personalized push messages for different users, intelligently pushes 5G messages for different types of mobile phones and different user preferences, and pushes concise text messages for the identified senior-friendly mobile phones.

[0010] S4. Perform real-time message push, monitor message delivery rate, establish delivery threshold, and simplify the 5G messages when the delivery rate is lower than the threshold.

[0011] Preferably, in step S1, the mobile phone information version and hardware configuration data are obtained from mobile phone manufacturers, e-commerce platforms, and suppliers; the intelligent classification is based on the characteristics of smartphones and feature phones to customize classification standards, and the mobile phones are classified as smartphones and feature phones based on different standards.

[0012] Preferably, the step of obtaining the user preference message type in step S2 is as follows:

[0013] Acquire user message reading behavior data, including message reading duration, message scrolling behavior, and whether the message was clicked to view;

[0014] Weight values ​​were assigned to different data features based on the analytic hierarchy process from reading behavior data.

[0015] The user's reading behavior value is calculated comprehensively, and the reading behavior value for different message types is calculated separately. The values ​​are then sorted, and the user with the highest reading behavior value is the message type that the user prefers.

[0016] Preferably, the weight values ​​are calculated by constructing a judgment matrix after building a hierarchical structure model, performing hierarchical sorting, and calculating the weights.

[0017] The specific steps for comprehensively calculating user reading behavior values ​​are as follows:

[0018] The user reading behavior value is obtained by multiplying the three data features of message reading time, message scrolling, and whether or not the message was clicked to view by their respective weights and then summing them up.

[0019] After calculating the reading behavior values ​​for all different types, the reading behavior values ​​are sorted from largest to smallest.

[0020] Preferably, the intelligent push 5G message step in step S3 is as follows:

[0021] Based on the classification results and the obtained user preference message types, a push strategy is formulated;

[0022] Analyze the historical behavior data of user groups to identify user groups with similar interests to the target users, and push the current 5G messages to user groups with similar interests;

[0023] Generate push time and push frequency strategies;

[0024] The push time is determined by analyzing the user's historical daily routine patterns, and a unique push time window is generated for each user.

[0025] The push frequency strategy is based on user feedback and engagement analysis. Users are categorized into active and inactive users. The push frequency is increased for active users and decreased for inactive users.

[0026] Preferably, the specific steps for obtaining the push time window are as follows:

[0027] The system acquires data on the time users spend on their mobile phones. Based on time series algorithms, it identifies patterns in users' mobile phone usage. There is a direct correlation between users' mobile phone usage time and their daily routines. In other words, the longer a user spends on their mobile phone, the more likely they are to be in a resting state. The system categorizes the daily routines of different users and pushes 5G messages to users based on their daily routines.

[0028] The specific steps for obtaining the push frequency strategy are as follows:

[0029] Establish an activity threshold, calculate user reading behavior values, and divide the reading behavior values ​​into different activity threshold ranges to obtain the user activity classification results. Increase the number of push notifications for users whose activity values ​​are above the activity threshold, and decrease the number of push notifications for users whose activity values ​​are below the activity threshold.

[0030] Preferably, in step S3, based on the results of intelligent classification, the user groups using senior-friendly mobile phones will have their messages simplified and then pushed to them.

[0031] Preferably, in step S4, the message push is based on WebSocket communication technology to build a push channel for 5G message push;

[0032] The monitoring SDK is integrated into the user's mobile phone. When a 5G message arrives on the phone, the SDK captures the message reception event and reports the message's unique identifier, reception time, mobile phone model, and network status information to the server.

[0033] When the client successfully receives the message, the SDK sends an acknowledgment to the server. The server marks the message status as "delivered" based on the message identifier in the acknowledgment. If no acknowledgment is received within a unit of time, the message status is marked as "pending acknowledgment".

[0034] For messages that fail to be sent multiple times, record the reasons for the failures, mark the message status as "not delivered", transfer the abnormal messages to the abnormal queue, and monitor the delivery rate of pushed messages.

[0035] Preferably, in step S4, a delivery threshold standard is established based on historical data, and the delivery rate of different users is compared with the threshold to determine whether it is lower than the delivery threshold.

[0036] Preferably, when the delivery rate is lower than the delivery threshold, message simplification is performed. This involves simplifying the content of the push message, simplifying the layout structure, and converting images into text to deliver the simplified 5G message to different users.

[0037] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0038] This invention establishes and categorizes mobile phone information databases to accurately match device-pushed content, improving effective reading rates; it analyzes reading feedback in real time to obtain user preferences, helping to generate personalized messages, enhancing user stickiness, optimizing the user experience for the elderly, and continuously collecting data and optimizing strategies to quickly adapt to changes in market and user needs, enabling the platform to maintain a competitive edge, provide high-quality services to users, accurately match users' current interests and needs, and increase user engagement. Attached Figure Description

[0039] Figure 1 This is a flowchart of the push steps of the present invention. Detailed Implementation

[0040] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0041] Reference Figure 1 As shown, a personalized 5G message push method based on intelligent agents includes the following push steps:

[0042] S1. Establish a mobile phone information database, obtain information on different mobile phone versions and hardware configuration data, and perform intelligent classification to divide mobile phones into smartphones and feature phones for the elderly.

[0043] S2. Obtain real-time reading feedback on different push messages, analyze and judge the user's preferred message types;

[0044] S3 generates personalized push messages for different users, intelligently pushes 5G messages for different types of mobile phones and different user preferences, and pushes concise text messages for the identified senior-friendly mobile phones.

[0045] S4. Perform real-time message push, monitor message delivery rate, establish delivery threshold, and simplify the 5G messages when the delivery rate is lower than the threshold.

[0046] During the mobile phone classification stage, the intelligent agent integrates hardware configuration and version data obtained from multiple channels (mobile phone manufacturers, e-commerce platforms, suppliers), establishes a classification model based on the core characteristics of smartphones and feature phones (such as screen size and functional complexity), and automatically completes the device type classification. This intelligent classification directly determines the content adaptation direction—matching rich media content (combining text, images, audio and video) for smartphone users and generating concise text content for feature phone users, thus achieving content and device adaptation from the source.

[0047] In the user preference analysis stage, the intelligent agent constructs a weight model based on the analytic hierarchy process to quantitatively analyze real-time feedback data such as user reading time, swiping behavior, and clicking actions. It automatically calculates and sorts the reading behavior values ​​of different message types, accurately locates the user's preferred content, and if the user has a high weighting for reading health information and clicks frequently, the intelligent agent can determine that the user prefers this type of content and prioritize pushes relevant 5G messages to achieve intelligent matching of content and user interests.

[0048] During the push strategy formulation phase, the intelligent agent integrates device classification results with user preferences to generate a dynamic push plan: by analyzing historical behavioral data, it identifies user groups with similar interests to expand the reach of content; based on time series algorithms, it mines user activity patterns to automatically generate exclusive push time windows (such as pushing news to office workers during their commute and pushing leisure content to retired users in the afternoon); and it dynamically adjusts the push frequency based on user activity, increasing the number of pushes of preferred content to active users and reducing interference to inactive users to ensure the timeliness and appropriateness of content pushes.

[0049] In message optimization and delivery monitoring, the intelligent agent monitors the push status in real time. When the delivery rate is lower than the threshold, it automatically triggers a content simplification mechanism—refining and extracting text content, simplifying the layout structure, and converting images into text descriptions to ensure that the content can adapt to scenarios with poor network environment or weak device performance. The intelligent agent continuously learns user feedback data and constantly optimizes the classification model, preference weights, and push strategy to ensure that content push always maintains a dynamic match with user needs, device capabilities, and scenario characteristics, ultimately realizing the personalized and intelligent upgrade of 5G messaging services.

[0050] In step S1, mobile phone information version and hardware configuration data are obtained from mobile phone manufacturers, e-commerce platforms, and suppliers; intelligent classification is based on the characteristics of smartphones and feature phones, and classification standards are customized to classify mobile phones into smartphones and feature phones based on different standards.

[0051] This application obtains data from multiple channels, including mobile phone manufacturers, e-commerce platforms, and suppliers, to comprehensively cover mobile phone information. Official data provided by manufacturers ensures the authority of versions and configurations, while user reviews and sales data from e-commerce platforms supplement actual usage. Supplier data reflects the characteristics of products circulating in the market. The mutual verification of multi-source data avoids the limitations of a single channel, making the mobile phone information database more comprehensive and accurate. This provides a reliable basis for subsequent intelligent classification and push strategy formulation, enhancing the scientific nature of decision-making.

[0052] Based on the characteristics of smartphones and feature phones, customized classification standards are developed, such as differentiation based on screen size, user interface, and functional complexity. This allows for precise segmentation of phone types, enabling the delivery of tailored messages to different user types. Smartphone users receive feature-rich 5G messages, while feature phone users receive concise text messages. This improves the matching degree between messages and user needs, avoids invalid push notifications, and makes push notifications more targeted, effectively enhancing user experience and message delivery.

[0053] The steps for obtaining user preference message types in step S2 are as follows:

[0054] Acquire user message reading behavior data, including message reading duration, message scrolling behavior, and whether the message was clicked to view;

[0055] Weight values ​​were assigned to different data features based on the analytic hierarchy process from reading behavior data.

[0056] The user's reading behavior value is calculated comprehensively, and the reading behavior value for different message types is calculated separately. The values ​​are then sorted, and the user with the highest reading behavior value is the message type that the user prefers.

[0057] The specific calculation formula is as follows:

[0058] Assume there is a total There are several types of messages, denoted as follows: For each type of message ( Define the following variables:

[0059] Let's assume reading the message The duration is ,use This indicates that if a message slides to a certain percentage, it is recorded as... Otherwise, record as ,use This indicates that clicking to view the message will be recorded as... Otherwise, record as ;

[0060] Based on the Analytic Hierarchy Process (AHP), let the weights for reading time, message scrolling, and whether or not the message was clicked be as follows: , , And satisfy , , , ;

[0061] Then for message types Reading behavior value The calculation formula is:

[0062]

[0063] right Sort the data and find the maximum value. ,Right now:

[0064]

[0065] like ( ), then message type This refers to the types of messages that users prefer.

[0066] The steps for intelligently pushing 5G messages in the S3 process are as follows:

[0067] Based on the classification results and the obtained user preference message types, a push strategy is formulated;

[0068] Analyze the historical behavior data of user groups to identify user groups with similar interests to the target users, and push the current 5G messages to user groups with similar interests;

[0069] Generate push time and push frequency strategies;

[0070] The push time is determined by analyzing the user's historical daily routine patterns, and a unique push time window is generated for each user.

[0071] The push frequency strategy is based on user feedback and engagement analysis. Users are categorized into active and inactive users. The push frequency is increased for active users and decreased for inactive users.

[0072] This application develops push strategies based on mobile phone classification and user preferences, enabling precise matching of 5G messages to corresponding groups. For example, it can push photography competition messages to photography enthusiasts who use high-pixel 5G phones, and push game information optimized for 5G networks to game enthusiasts. This avoids ineffective pushes caused by "casting a wide net" of messages, ensuring that message content is highly aligned with user needs, significantly improving message open rates, click-through rates, and conversion rates, and maximizing the value of each push.

[0073] By analyzing historical behavioral data, we can identify user groups with similar interests and push notifications to them. This allows us to discover potential users who are interested and generate personalized push notification windows based on users' historical daily routines, taking into full account users' lifestyles and usage habits.

[0074] The specific steps for obtaining the push time window are as follows:

[0075] The system acquires data on the time users spend on their mobile phones. Based on time series algorithms, it identifies patterns in users' mobile phone usage. There is a direct correlation between users' mobile phone usage time and their daily routines. In other words, the longer a user spends on their mobile phone, the more likely they are to be in a resting state. The system categorizes the daily routines of different users and pushes 5G messages to users based on their daily routines.

[0076] The specific steps for obtaining the push frequency strategy are as follows:

[0077] Establish an activity threshold, calculate user reading behavior values, and divide the reading behavior values ​​into different activity threshold ranges to obtain the user activity classification results. Increase the number of push notifications for users whose activity values ​​are above the activity threshold, and decrease the number of push notifications for users whose activity values ​​are below the activity threshold.

[0078] In step S3, based on the results of intelligent classification, messages for users of older mobile phones are simplified and then pushed to them.

[0079] The specific calculation steps are as follows:

[0080] Steps to obtain the push time window:

[0081] Set user The set of timestamps for a day is ,in This indicates the specific time points and durations during which a user uses their phone. , For timestamp Corresponding mobile phone usage time;

[0082] Calculate the total usage time of the user's mobile phone. Total daily mobile phone usage time The calculation formula is:

[0083]

[0084] Based on time series algorithms to identify usage patterns, assuming a seasonal decomposition algorithm is used, the time series... Decomposed into trend terms Seasonal items and remainder By analyzing trend and seasonality factors, user usage patterns can be identified.

[0085] Classify daily routines and set thresholds for each category. ,like Determine if the user is active; It determines whether a user is in a resting state and categorizes users into different sleep / rest schedules. Active type, Rest type ;

[0086] Generate push notification time windows based on user's daily routine. For users Generate push time window For active users, the push notification window is set during their active periods; for resting users, the push notification window is set to avoid their resting periods.

[0087] The steps to obtain the push frequency strategy are as follows:

[0088] Set user Reading behavior value (Calculated based on the previous formula), the set of activity thresholds is as follows: Divide users into Each activity category

[0089] User activity classification Activity categories The calculation method is as follows:

[0090]

[0091] Set user The initial number of pushes is Adjusted push notification frequency For: a certain low activity threshold, a medium activity threshold, a certain high activity threshold, and a certain high activity threshold.

[0092]

[0093] in To reduce the coefficient of push notification frequency, To increase the number of push notifications;

[0094] Message Simplification Based on Intelligent Classification

[0095] Set user The mobile phone classification results are Smartphones and phones for seniors Regarding the message content ,like Older phones, on the other hand, handle messages... The simplified message is obtained by performing simplification processing. This can be represented as: senior-friendly mobile phone or smartphone.

[0096]

[0097] in Indicates the message A function for simplifying content and optimizing formatting.

[0098] In step S4, message push is based on WebSocket communication technology to build a push channel for 5G message push.

[0099] The monitoring SDK is integrated into the user's mobile phone. When a 5G message arrives on the phone, the SDK captures the message reception event and reports the message's unique identifier, reception time, mobile phone model, and network status information to the server.

[0100] When the client successfully receives the message, the SDK sends an acknowledgment to the server. The server marks the message status as "delivered" based on the message identifier in the acknowledgment. If no acknowledgment is received within a unit of time, the message status is marked as "pending acknowledgment".

[0101] For messages that fail to be sent multiple times, record the reasons for the failures, mark the message status as "not delivered", transfer the abnormal messages to the abnormal queue, and monitor the delivery rate of pushed messages.

[0102] This application uses WebSocket communication technology to build a push channel, which can establish a persistent bidirectional communication connection between the client and the server, breaking the limitations of the traditional HTTP request-response mode. Compared with the traditional method, WebSocket can realize real-time message push without frequent connection establishment and disconnection, which greatly reduces communication latency. The monitoring SDK is integrated into the user's mobile phone to accurately capture message receiving events and report key information such as message unique identifier and receiving time.

[0103] For messages that fail to be sent multiple times, record the reasons for the failure in detail and transfer them to the exception queue to facilitate technicians in quickly locating and resolving the problem.

[0104] In step S4, a delivery threshold standard is established based on historical data. The delivery rate of different users is compared with the threshold to determine whether it is lower than the delivery threshold. If the delivery rate is lower than the delivery threshold, message simplification is performed. By simplifying the content of the push message, simplifying the layout structure, and converting images into text, the simplified 5G message is delivered to different users.

[0105] This application establishes a delivery threshold based on historical data, quantifies and analyzes past message delivery status, user device, and network environment information, making the threshold setting more closely aligned with actual scenarios. When the user delivery rate falls below the threshold, a simplification mechanism is triggered, realizing a shift from "experience-driven" to "data-driven." The simplified messages are more in line with user receiving habits. For users with older or low-configuration devices, concise text messages reduce the reading burden; for users with poor network conditions, fast-loading messages reduce waiting time.

[0106] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention.

Claims

1. A 5G personalized message push method based on intelligent agents, characterized in that, The push notification steps are as follows: S1. Establish a mobile phone information database, obtain information on different mobile phone versions and hardware configuration data, and perform intelligent classification to divide mobile phones into smartphones and feature phones for the elderly. S2. Obtain real-time reading feedback on different push messages, analyze and judge the user's preferred message types; S3 generates personalized push messages for different users, intelligently pushes 5G messages for different types of mobile phones and different user preferences, and pushes concise text messages for the identified senior-friendly mobile phones. S4. Perform real-time message push, monitor message delivery rate, establish delivery threshold, and simplify the 5G messages when the delivery rate is lower than the threshold. In step S4, message push is based on WebSocket communication technology to build a push channel for 5G message push. The monitoring SDK is integrated into the user's mobile phone. When a 5G message arrives on the phone, the SDK captures the message reception event and reports the message's unique identifier, reception time, mobile phone model, and network status information to the server. When the client successfully receives the message, the SDK sends an acknowledgment to the server. The server marks the message status as "delivered" based on the message identifier in the acknowledgment. If no acknowledgment is received within a unit of time, the message status is marked as "pending acknowledgment". For messages that fail to be sent multiple times, record the reason for the failure, mark the message status as "not delivered", transfer the abnormal message to the abnormal queue, and monitor the delivery rate of the pushed messages.

2. The 5G message personalized push method based on intelligent agents according to claim 1, characterized in that, In step S1, mobile phone information version and hardware configuration data are obtained from mobile phone manufacturers, e-commerce platforms, and suppliers; intelligent classification is based on the characteristics of smartphones and feature phones, and classification standards are customized to classify mobile phones into smartphones and feature phones based on different standards.

3. The 5G message personalized push method based on intelligent agents according to claim 1, characterized in that, The steps for obtaining user preference message types in step S2 are as follows: Acquire user message reading behavior data, including message reading time, message scrolling, and whether the message was clicked to view; Weight values ​​were assigned to different data features based on the analytic hierarchy process from reading behavior data. The system comprehensively calculates user reading behavior values, calculates reading behavior values ​​for different message types separately, sorts them, and obtains the user reading behavior value with the highest ranking. The user reading behavior value with the highest ranking is the message type that the user prefers.

4. The 5G message personalized push method based on intelligent agents according to claim 3, characterized in that, The Analytic Hierarchy Process (AHP) calculates weight values ​​by constructing a hierarchical structure model, building a judgment matrix, sorting the hierarchy, and calculating the weights. The specific steps for comprehensively calculating user reading behavior values ​​are as follows: The user reading behavior value is obtained by multiplying the three data features of message reading time, message scrolling, and whether or not the message was clicked to view by their respective weights and then summing them up. After calculating the reading behavior values ​​for all different types, the reading behavior values ​​are sorted from largest to smallest.

5. The 5G message personalized push method based on intelligent agents according to claim 1, characterized in that, The steps for intelligently pushing 5G messages in the S3 process are as follows: Based on the classification results and the obtained user preference message types, a push strategy is formulated; Analyze the historical behavior data of user groups to identify user groups with similar interests to the target users, and push the current 5G message to user groups with similar interests; Generate push time strategy and push frequency strategy; The push time is determined by analyzing the user's historical daily routine patterns, and a unique push time window is generated for each user. The push frequency strategy is based on user feedback and engagement analysis. Users are categorized into active and inactive users. The push frequency is increased for active users and decreased for inactive users.

6. The 5G message personalized push method based on intelligent agents according to claim 5, characterized in that, The specific steps for obtaining the push time window are as follows: The system acquires data on the time users spend on their mobile phones. Based on time series algorithms, it identifies patterns in users' mobile phone usage. There is a direct correlation between users' mobile phone usage time and their daily routines. The longer the user spends on their mobile phone, the more likely they are to be in a resting state. The system categorizes the daily routines of different users and pushes 5G messages to users based on their daily routines. The specific steps for obtaining the push frequency strategy are as follows: Establish an activity threshold, calculate user reading behavior values, and divide the reading behavior values ​​into different activity threshold ranges to obtain the user activity classification results. Increase the number of push notifications for users whose activity values ​​are above the activity threshold, and decrease the number of push notifications for users whose activity values ​​are below the activity threshold.

7. The 5G message personalized push method based on intelligent agents according to claim 1, characterized in that, In step S3, based on the results of intelligent classification, messages for users of older mobile phones are simplified and then pushed to them.

8. The 5G message personalized push method based on intelligent agents according to claim 1, characterized in that, In step S4, a delivery threshold standard is established based on historical data. The delivery rate of different users is compared with the threshold to determine whether it is below the delivery threshold.

9. A 5G message personalized push method based on intelligent agents according to claim 8, characterized in that, When the delivery rate is lower than the delivery threshold, the message is simplified. This involves streamlining the content of the push message, simplifying the layout, and converting images into text to deliver the simplified 5G message to different users.

Images