Information Push Method Using User Behavior Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing information push methods have a limited coverage range and low precision due to reliance on popularity-based and interest-type-based systems, which fail to account for user-specific preferences and lead to information loss.
Innovation Solution
An information push method that analyzes user-behavioral data to determine second feature information of a target object, calculates user-preference levels based on this data and the feature information of to-be-pushed content, and prioritizes information accordingly to enhance relevance and coverage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If popularity-based pushing is used to rank information according to statistics such as attention degree and quantity of good comments, then high-quality information can be ranked at prior push positions, but user-specific pushing cannot be performed and the coverage range is limited
Solution Approach 1:
The patent segments the information push system into multiple independent modules: a user behavior analysis module that processes individual user data, a feature extraction module that identifies user preferences, and a push decision module that combines popularity metrics with user-specific features. This segmentation enables both global quality ranking and individualized pushing simultaneously
Solution Approach 2:
The patent adds a new dimension to the information push system by introducing user behavior feature vectors alongside traditional popularity metrics. Instead of relying solely on scalar popularity scores, the system now operates in a multi-dimensional space that includes user preferences, browsing history, and interaction patterns, enabling more nuanced and adaptable information delivery
2Adaptability or versatility
If interest-type-based pushing is used to obtain information types a target object is interested in based on browsing and use history, then preferential pushing can be performed, but information loss in type system limits the coverage range
Solution Approach 1:
The patent transforms the rigid type system into a flexible parameter-based system. Instead of categorizing information into fixed types, the system extracts continuous feature parameters from user behavior data (such as engagement duration, interaction frequency, and preference强度) and uses these parameters to dynamically adjust information delivery, thereby reducing information loss while maintaining adaptability
Solution Approach 2:
The patent creates a composite information delivery system that combines multiple data sources: user browsing history, interaction patterns, content features, and popularity metrics. This composite approach integrates diverse information types into a unified push decision framework, expanding coverage beyond what any single data source could provide while preserving user preferences
Data Source
AI summary
Embodiments of this application provide an information push method performed at a computing device. The computing device determines second feature information of a target object associated with a terminal that is communicatively connected to the computing device according to user-behavioral data of the target object and first feature information of to-be-pushed information, and then user-preference level information according to the second feature information of the target object and the first feature information of the to-be-pushed information. The computing device then orders the to-be-pushed information according to the user-preference level information into target information. Finally, the computing device pushes the target information to the target object. Therefore, according to the technical solutions provided in the embodiments of this application, a coverage range of information can be expanded to some extent, thereby improving the preciseness of information pushing.