Web Page Ranking Updates Using Time and Click Signals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current search indices assign static rankings to web pages at the time of discovery, which can become outdated due to changes in web page quality over time, leading to inaccurate search result recommendations.
Innovation Solution
Implement a time and click-based updatable static ranking system that uses two ranking models to continuously evaluate and re-rank web pages based on features available at discovery and post-discovery, incorporating metrics such as click data and time elapsed since discovery.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If static ranking is assigned at discovery time, then device complexity is reduced and processing speed is improved, but ranking accuracy deteriorates over time due to outdated web page quality information
Solution Approach 1:
The patent transforms the static ranking system into a dynamic one by implementing continuous re-ranking at predetermined intervals. The search index automatically updates web page rankings based on newly available features and data without requiring full re-crawling, allowing the system to adapt to changing web page quality over time while maintaining efficient operation.
Solution Approach 2:
The system performs preliminary actions by pre-scheduling re-ranking operations at predetermined time intervals. This allows the search index to proactively update rankings before accuracy degrades significantly, ensuring that web page quality changes are reflected timely without requiring reactive full-reindexing operations.
2Measurement precision
If full re-crawling and re-indexing is performed to update rankings, then ranking accuracy is improved, but productivity decreases due to high computational cost and time consumption
Solution Approach 1:
The patent segments the re-ranking process into independent, incremental operations that focus only on updating rankings based on new features rather than re-processing entire web pages. This allows the system to maintain accurate rankings without the computational overhead of full re-crawling and re-indexing, preserving both accuracy and productivity.
Solution Approach 2:
The system updates rankings by changing specific parameters (newly available features) rather than re-evaluating all parameters from scratch. This selective parameter update approach reduces computational complexity while maintaining ranking accuracy, allowing fast updates without full re-processing.
3Adaptability or versatility
If multiple ranking models are used for continuous evaluation, then adaptability to web page quality changes is improved, but device complexity increases
Solution Approach 1:
The patent merges multiple ranking model evaluations into a unified search index update process. Instead of maintaining separate complex systems for different ranking models, the system integrates their outputs into a single continuous re-ranking mechanism that updates the search index atomically, reducing overall system complexity while maintaining adaptability.
Solution Approach 2:
The search index structure is designed to be universal and multi-functional, supporting both initial ranking at discovery time and continuous re-ranking at predetermined intervals. This unified structure handles multiple ranking scenarios without requiring separate specialized systems, reducing complexity while maintaining versatility.
Data Source
AI summary
Aspects of the disclosure include methods and systems for a time and click based updatable static web page ranking. An exemplary method includes identifying a web page at a discovery time, predicting, by a first ranking model, a first ranking of the web page using first features available at the discovery time, and predicting, by a second ranking model, a second ranking of the web page using the first features. An index score is generated from the first ranking and the second ranking. The method includes, after discovery time, predicting, by the first ranking model, an updated first ranking of the web page using the first features, predicting, by the second ranking model, an updated second ranking of the web page using the first features and second features not available at the discovery time, and updating the index score from a combination of the updated first ranking and the updated second ranking.