Search Result Reordering Using Real-Time User Interaction Signals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing search systems struggle with accurately presenting relevant search results due to the intermixing of different types of media content, leading to increased time spent by users finding desired content and poor search efficiency.
Innovation Solution
A method and apparatus that reorder search results based on real-time interactive behaviors of users, using a reordering model trained on historical and real-time data to adjust content sequences dynamically, incorporating both short-term and long-term user interests.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If search results are presented in a fixed order based on initial matching, then the system structure is simple and easy to implement, but the search efficiency deteriorates because users cannot quickly find desired content amidst intermixed different types of videos
Solution Approach 1:
The patent implements dynamic reordering of search results based on real-time user interactive behaviors. The search result list is no longer static but dynamically adjusted according to user actions such as playback completion, liking, sharing, and commenting. This dynamic approach resolves the contradiction by making the system adaptive to user needs while maintaining reasonable implementation complexity through modular design.
Solution Approach 2:
The system incorporates feedback mechanisms by continuously monitoring user interactive behaviors and using this feedback to reorder search results. The reordering model receives feedback from user actions and adjusts the presentation order accordingly, thereby improving search efficiency without requiring completely complex system restructuring.
2Measurement precision
If the search result page displays all media contents in the initial order, then the presentation is simple and fast, but the accuracy of search result presentation deteriorates because different types of videos are intermixed and do not align with user intent
Solution Approach 1:
The system performs preliminary actions by pre-training a reordering model on historical user behavior data before actual search operations. This pre-trained model is ready to quickly reorder results based on real-time interactions, thereby improving presentation accuracy without adding significant time delays during the actual search process.
Solution Approach 2:
The search result presentation transitions from a static initial order to a dynamic order that adapts to user interactions. The system dynamically reorders results based on real-time behavioral signals, improving accuracy while minimizing time loss through efficient real-time processing of user actions.
3Ease of operation
If the search result order remains static after initial display, then the system operation is simple and fast, but user satisfaction deteriorates because the content order does not align with evolving user intent during the search session
Solution Approach 1:
The system implements self-service by automatically reordering search results based on user behaviors without requiring manual user intervention. The reordering model autonomously processes user actions and adjusts the result order, maintaining ease of operation while improving user satisfaction through adaptive content presentation.
Solution Approach 2:
The system uses feedback from user interactive behaviors to continuously improve result presentation. By monitoring user actions and incorporating this feedback into the reordering model, the system maintains simple operation while reliably adapting to user intent, thereby improving satisfaction.
Data Source
AI summary
A computer device performs a search result reordering method. The computer device displays a search result page corresponding to a search session, the search result page displaying a first content sequence obtained based on a search keyword corresponding to the search session, and the first content sequence including a plurality of media contents. After obtaining at least one real-time interactive behavior performed by a target object on the search result page, the computer device determines media contents other than a target media content associated with the at least one real-time interactive behavior as to-be-sorted media contents in the first content sequence; and reorders the to-be-sorted media contents to obtain a second content sequence when the at least one real-time interactive behavior satisfies a reordering trigger condition, and updates the search result page based on the second content sequence.