Query Correction Using Derivative Search and Resource Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing search systems face difficulties in correcting queries with potentially inaccurate terms, such as unique, rarely occurring, class-instance, and misspelled terms, as traditional query correction processes rely on historical analysis and may provide inadequate corrections.
Innovation Solution
The method involves receiving an original query, identifying potentially inaccurate terms, generating derivative queries without these terms, submitting them for search, and selecting corrected terms based on responsive resources to generate a corrected query, which is then used for search operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional query correction processes relying on historical analysis are used, then the correction process is simple, but the correction accuracy is insufficient for unique, rarely occurring, class-instance, and misspelled terms
Solution Approach 1:
The query correction process is segmented into multiple independent stages: (1) generating derivative queries by removing potentially inaccurate terms, (2) searching for resources responsive to each derivative query, (3) analyzing co-occurring terms in the resources, (4) scoring candidate correction terms based on multiple criteria, and (5) selecting the best correction. This segmentation allows each stage to focus on a specific aspect of correction accuracy without overwhelming complexity.
Solution Approach 2:
The patent introduces intermediary elements including derivative queries (which serve as mediators between the original query and correction candidates), resource corpora (as intermediaries providing contextual evidence), and scoring functions (as mediators evaluating candidate quality). These intermediaries enable accurate correction of difficult terms by providing additional contextual information without requiring direct complex analysis of the original query alone.
2Reliability
If derivative queries are generated and searched independently without relying on original query resources, then a larger and higher-quality set of resources is available for correction, but the processing time and computational resources increase
Solution Approach 1:
The system performs preliminary actions by generating derivative queries that exclude potentially inaccurate terms before conducting the main correction analysis. This preliminary step creates a cleaner search foundation that returns higher-quality resources, making the subsequent correction process more reliable. The derivative queries are prepared in advance to establish a solid basis for accurate term correction.
Solution Approach 2:
The patent changes key parameters of the search process by modifying query composition (removing potentially inaccurate terms to create derivative queries) and adjusting scoring criteria (using multiple factors including term frequency, document frequency, and positional information). These parameter changes enable the system to leverage larger resource sets effectively while maintaining efficient processing through optimized scoring mechanisms.
3Measurement precision
If multiple candidate correction terms are evaluated with independent resource identification, then the quality of correction analysis improves, but the system complexity and resource requirements increase
Solution Approach 1:
The scoring function is designed as a universal multi-functional mechanism that evaluates candidate correction terms using multiple criteria simultaneously: term frequency in resources, document frequency across the corpus, positional information from search results, and co-occurring term analysis. This universal scoring approach handles diverse correction scenarios (misspellings, class-instance errors, rare terms) with a single integrated system rather than separate specialized processes.
Solution Approach 2:
The system implements feedback mechanisms where resources identified from derivative queries provide evidence that feeds back into evaluating candidate correction terms. The scoring function continuously refines candidate rankings based on feedback from resource analysis, including term co-occurrence patterns and positional information. This iterative feedback process improves correction precision while managing system complexity through structured feedback loops rather than uncontrolled system expansion.
Data Source
AI summary
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for correcting potentially inaccurate terms in queries. In an aspect, a search system determines that an original query includes a potentially inaccurate term. The determination is made on the basis that the term of the original query meets an inaccuracy criterion. Derivative queries are generated from the original query. Each derivative query does not include the potentially inaccurate term. For each derivative query resources that are determined to be responsive to the derivative query are determined, and from the resources, a corrected term is determined. The corrected term is substituted for the potentially inaccurate term, and a search operation that uses the corrected query as input is performed.