Query Burst Detection for Event-Related Information Presentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional search engines fail to detect short-term trends or current events due to their reliance on absolute numbers of searches over long periods, making it difficult to identify sudden increases in query interest that do not significantly impact long-term popularity listings.
Innovation Solution
A system that detects bursts in query entries by tracking query rates over time and presents event-related information when a sudden increase exceeds a threshold rate, allowing users to identify causes for the surge and plan inventory accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If search engines use absolute numbers of searches over long periods to determine popularity, then long-term trends are accurately reflected, but short-term trends and sudden query increases are not detected
Solution Approach 1:
The patent segments the analysis of search query popularity into different time windows. Instead of using a single long-term absolute count, the system divides queries into recent queries (short-term window) and historical queries (long-term window), allowing separate analysis of short-term trends and long-term popularity. This segmentation enables detection of sudden query increases while maintaining context of overall popularity.
Solution Approach 2:
The system dynamically adjusts the weighting between recent and historical query counts based on detected patterns. When a sudden increase in queries is detected (burst detection), the system increases the weight of recent queries to highlight the short-term trend. This dynamic adjustment allows the popularity metric to adapt to changing query patterns in real-time.
2Quantity of substance
If search engines rely on absolute search numbers, then overall popularity is accurately measured, but sudden increases in query interest are not identified
Solution Approach 1:
The system implements feedback by continuously monitoring the rate of query submissions and comparing recent rates against historical baselines. When the query rate exceeds a threshold or shows significant acceleration, the system generates a burst signal that triggers updated popularity rankings. This feedback loop enables detection of sudden query increases that would be invisible in absolute counts alone.
Solution Approach 2:
The patent changes the parameter from absolute query count to query rate (queries per time unit). By measuring the velocity of query submissions rather than just the cumulative total, the system can detect sudden increases in interest. The query rate parameter is calculated as the change in query count divided by the change in time, enabling identification of bursts even when total volumes remain moderate.
3Productivity
If search engines update popularity listings frequently, then current trends are reflected, but system complexity and computational resources increase
Solution Approach 1:
The system uses periodic action by updating popularity rankings at scheduled intervals (e.g., hourly, daily) rather than continuously. Between updates, the system accumulates query data and detects bursts. This periodic approach balances responsiveness with computational efficiency, as full ranking recalculations are performed only when necessary rather than continuously.
Solution Approach 2:
The patent introduces an intermediary burst detection mechanism that sits between raw query data and the popularity ranking system. This intermediary layer pre-processes query streams to identify bursts and generates simplified signals for the ranking system. This mediation reduces the computational burden on the main ranking algorithm by filtering out normal variations and only triggering updates when significant changes occur.
Data Source
AI summary
In an embodiment, a method is provided for presenting a query directed at an information resource. In this method, a number of queries is accessed over a time period. A burst of the number of queries is detected within the time period. It should be noted that a burst is a determinable increase in the number of queries received within the time period relative to a historical number of queries received in a preceding time interval. Event-related information that is associated with the burst in the time is searched, and the query in conjunction with the event-related information is displayed at a display unit.