Information Retrieval System for Reducing Overly Broad Query Execution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional information retrieval systems face inefficiencies due to users entering overly broad queries, leading to excessive system resource consumption and failure to converge on desired information, particularly for users lacking subject-matter expertise.
Innovation Solution
An information retrieval system that detects overly broad queries and presents users with a topical classification system, including popular documents and their counts, allowing users to navigate and refine their queries without executing the query against the database, thereby reducing unnecessary searches and improving information retrieval efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If users enter overly broad queries to explore subject areas, then users can gain insights into unfamiliar topics, but system resources are consumed excessively due to large result sets
Solution Approach 1:
The patent segments the broad query results into classified topic categories with representative documents. Instead of presenting all documents from a broad query, the system divides them into organized topics, allowing users to explore subject areas efficiently while reducing the display burden and system resource consumption.
Solution Approach 2:
The patent introduces an intermediary classification layer between the user's broad query and the actual document results. This intermediary system categorizes documents by topic and presents representative samples, mediating between the user's exploration needs and system resource constraints.
2Measurement precision
If users iteratively enter more targeted queries to converge on desired information, then search precision improves, but the number of queries and system resource usage increase
Solution Approach 1:
The patent performs preliminary classification and organization of documents by topic before the user submits their query. When a broad query is entered, the system has already prepared categorized results with representative documents, allowing users to immediately see relevant topics without needing to iteratively refine their queries multiple times.
Solution Approach 2:
The patent provides feedback to users by displaying classified topics and representative documents from broad queries. This feedback mechanism helps users understand the available information structure and makes informed decisions about which topics to explore, reducing the need for multiple iterative queries.
3Productivity
If conventional systems add search engine capacity to handle large numbers of queries, then query processing capability increases, but cost increases and inefficiency of overly broad queries is not addressed
Solution Approach 1:
The patent extracts and presents only the most relevant representative documents from broad query results, rather than processing and displaying all documents. This extraction approach reduces the processing burden on search engines while still providing users with valuable information from unfamiliar subject areas.
Solution Approach 2:
The patent uses a lightweight classification and representative document selection mechanism that does not require substantial search engine capacity. Instead of investing in expensive high-capacity search infrastructure, the system uses a simpler, more efficient approach that handles broad queries without requiring proportional increases in search engine capacity.
Data Source
AI summary
The present inventor devised, among other things, an exemplary information retrieval system that promises to reduce the execution of overly broad queries. One exemplary system detects overly broad queries and presents users one or more potentially relevant portions of a hierarchical subject matter classification system, instead of executing the query against the targeted database. The system also presents users the option of accessing one or more relevant documents to the query via an interface for the classification system.