Analogous Query Extraction via Per-User Search Logs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional search services fail to provide relevant search queries when users are unsure of suitable keywords, and they do not offer queries with similar attributes that do not contain the user's search query or its synonyms, leading to inefficiencies in information retrieval.
Innovation Solution
An analogous query extracting system and method that uses search log information to create a per-user database of search queries and query pairs, calculating conditional probabilities to identify and provide queries with similar attributes, even if they do not contain the original search query or its synonyms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional search services collect search queries containing user's search query or its synonym, then they can provide extended search keywords, but they fail to provide queries with similar attributes that do not contain the original search query
Solution Approach 1:
The patent transforms the approach from exact string matching (containing search query or synonym) to attribute-based matching by extracting and comparing query attributes. This parameter change enables the system to identify queries with similar attributes even when they don't contain the original search terms, thereby providing more versatile query suggestions without losing attribute-based information.
Solution Approach 2:
The patent introduces query attributes as an intermediary layer between the original search query and the suggested queries. Instead of directly matching search queries, the system extracts attributes from queries and uses these attributes as mediators to find and suggest related queries, enabling broader query coverage while maintaining relevance.
2Ease of operation
If service operators manually sort related search queries, then they can provide customized suggestions, but it causes temporal and financial burdens
Solution Approach 1:
The patent implements a self-service system where the automated query processing module automatically extracts query attributes, identifies related queries, and generates suggestions without requiring manual sorting by service operators. This automation eliminates the temporal and financial burdens of manual processing while maintaining customized query suggestions tailored to user needs.
Solution Approach 2:
The patent replaces the mechanical manual sorting process with an automated computational system that uses attribute extraction and comparison algorithms. This substitution eliminates human labor requirements while achieving the same or better customization效果, significantly reducing time and resource consumption.
3Reliability
If conventional services assume users know suitable search queries, then they can provide extended keywords, but they fail when users do not know suitable search queries for a search target
Solution Approach 1:
The patent performs preliminary attribute extraction and query analysis before the user needs search suggestions. By pre-processing queries to extract attributes and identifying potential related queries in advance, the system is prepared to provide guidance even when users don't know what to search for, improving both reliability and adaptability.
Solution Approach 2:
The patent implements a feedback mechanism where the system analyzes user search behavior and query attributes, then provides suggested queries that feed back to users. This feedback loop enables the system to learn from user interactions and improve its ability to guide users who are unsure about search terms, enhancing both reliability and user support.
Data Source
AI summary
A system for extracting analogous queries is disclosed. The system includes a search query extracting module, a query pair extracting module, and an analogous query extracting module. The search query extracting module creates a per-user search query DB by reading search queries inputted by users from a search log DB that contains user device identifiers, search queries inputted from user devices, and information about time when the users request a search. The query pair extracting module creates a per-user query pair DB by extracting, from the per-user search query DB, permutations of different two search queries among search queries inputted by a specific user. The analogous query extracting module reads query pairs having a specific query in the per-user query pair DB, and creates analogous queries by extracting counterparts of a specific query from the query pairs. The analogous queries are provided in response to a search request for a specific query from the user device.