Message Pushing Using Forecast Model Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current message pushing technologies lack efficiency as they send messages to all registered users without screening, resulting in inefficiency and resource waste, especially since some users may be offline, and do not accurately target active users for promotions.
Innovation Solution
A method and device for message pushing that utilize a forecast model, such as the C5.0 decision tree algorithm, to identify and select messages for specific target user groups based on user attributes like daily game rounds and login times, ensuring messages are only sent to active users, thereby improving resource utilization and promotional accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If messages are sent to all registered users without screening, then message coverage is maximized, but resource efficiency deteriorates and offline users are wasted
Solution Approach 1:
The system performs preliminary classification of users into active and inactive groups before message pushing. By pre-screening users based on their activity status, the system avoids wasting resources on offline users while ensuring active users receive messages, thus resolving the contradiction between coverage and efficiency
Solution Approach 2:
The user base is segmented into different groups (active users and inactive users) based on their online status. This segmentation allows the system to apply different message pushing strategies to different groups, sending messages only to active users to improve resource efficiency while maintaining effective coverage
2Quantity of substance
If messages are sent to all registered users, then promotional reach is maximized, but message precision deteriorates due to lack of user screening
Solution Approach 1:
The system performs preliminary classification of users into active and inactive groups before message pushing. By pre-screening users based on their activity status, the system avoids wasting resources on offline users while ensuring active users receive messages, thus resolving the contradiction between coverage and efficiency
Solution Approach 2:
The user base is segmented into different groups (active users and inactive users) based on their online status. This segmentation allows the system to apply different message pushing strategies to different groups, sending messages only to active users to improve resource efficiency while maintaining effective coverage
3Speed
If full push approach is used after message screening, then message delivery speed is maintained, but resource waste increases due to sending to offline users
Solution Approach 1:
The system performs preliminary classification of users into active and inactive groups before message pushing. By pre-screening users based on their activity status, the system avoids wasting resources on offline users while ensuring active users receive messages, thus resolving the contradiction between coverage and efficiency
Solution Approach 2:
Instead of performing a full push to all users, the system applies partial action by sending messages only to the screened active user group. This partial approach eliminates waste on offline users while maintaining delivery speed for the target audience
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Systems and methods are provided for pushing messages. For example, upon receiving a push trigger message, one or more first user identifiers corresponding to the push trigger message are acquired; in response to the acquired first user identifiers being in a predetermined target user group, one or more first push messages are selected based on at least information associated with a first promotion type of the target user group; and the first push messages are pushed based on at least information associated with the first user identifiers, wherein the target user group includes a plurality of second user identifiers and is generated by statistical processing of first user attributes of registered user identifiers based on at least information associated with a predetermined first forecast model.