Query Suggestion System Using Predicate Overlap Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Search engines often fail to provide adequate query suggestions for infrequently used queries and may experience topic drift, leading to irrelevant suggestions.
Innovation Solution
Analyzing query logs to identify sibling queries by determining the overlap in predicate queries, calculating a query map value or intersect frequency, and providing suggestions when the value exceeds a threshold, thereby minimizing topic drift and suggesting relevant queries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If query logs are analyzed using traditional methods, then query suggestions can be provided for frequently used queries, but adequate suggestions cannot be provided for infrequently used queries
Solution Approach 1:
The patent segments the query analysis process into multiple components: identifying predicate queries, calculating query map values, determining intersect frequencies, and applying multiple thresholds. This segmentation allows the system to handle both frequent and infrequent queries by breaking down the complex analysis into manageable steps that can be applied uniformly across all query types
Solution Approach 2:
The patent introduces multiple parameters including query map value, intersect frequency, and various thresholds to quantify query relationships. By changing from simple frequency-based suggestions to multi-parameter analysis, the system can reliably identify suggestions for infrequently used queries based on their contextual relationships with predicate queries rather than just raw frequency
2Productivity
If query suggestions are provided based on previous queries in query logs, then suggestions can be generated, but topic drift occurs leading to irrelevant suggestions
Solution Approach 1:
The patent implements feedback mechanisms by calculating query map values that measure the relationship between current queries and predicate queries. The system uses intersect frequency as feedback to determine whether suggested queries maintain topic relevance, allowing it to generate suggestions while preventing topic drift through quantitative relevance assessment
Solution Approach 2:
The patent replaces simple mechanical frequency-based suggestion generation with a more sophisticated system that uses calculated metrics (query map values, intersect frequencies) to determine suggestion relevance. This substitution allows the system to maintain productivity while improving reliability by using quantitative measures of topical relatedness
Data Source
AI summary
Methods, systems, and apparatus, including computer program products, in which one or more search query suggestions can be identified based upon the intersection of predicate search groups. A first predicate search group can be compared to a second predicate search group to derive a query map value associated with the intersection of the predicate search groups. The query map value can be compared to a threshold to determine whether to identify different search queries as suggestions for one another.