Guided Search Query Model Similarity Ranking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing guided search systems are limited in providing guidance information when the user's query does not exactly match the query information stored in the mapping table, failing to retrieve relevant information if there is no exact match.
Innovation Solution
The guided search server derives models from the user's query information using natural language processing techniques, calculates similarity scores, and ranks these models to retrieve guidance information from a database, even if the query is not an exact match, by using the highest-ranked models as keywords to search for corresponding guidance information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the guided search server uses exact matching to retrieve guidance information from the mapping table, then the retrieval precision is high, but the system cannot provide guidance information for queries that do not exactly match stored entries
Solution Approach 1:
The patent transforms the exact matching parameter into a similarity scoring parameter. Instead of requiring queries to exactly match stored entries, the system calculates similarity scores between queries and stored information, allowing flexible matching while maintaining retrieval precision through score-based ranking.
Solution Approach 2:
The patent introduces models as intermediary representations between user queries and the mapping table. These models capture the semantic meaning of queries and enable comparison with stored information, serving as a bridge that allows flexible yet precise information retrieval without requiring exact string matches.
2Adaptability or versatility
If the guided search server stores a comprehensive mapping table to cover all possible queries, then the coverage is improved, but the system complexity and storage requirements increase
Solution Approach 1:
The patent creates simplified model representations (copies) of the essential semantic information from queries and stored data. These model copies enable comparison and matching without requiring the system to store and process all possible original query variations, reducing complexity while maintaining coverage.
Solution Approach 2:
The patent designs the mapping table to store universal model representations that can match multiple different query formulations. A single stored entry can serve as a match for various queries with similar meanings, reducing the total number of entries needed while expanding the effective coverage.
Data Source
AI summary
Providing guided search includes: receiving a new query; processing the new query to obtain a plurality of models, wherein a model indicates a mapping relationship of a previously stored query and corresponding guidance information; determining a corresponding plurality of similarities of the plurality of models relative to the new query; obtaining guidance information in a database comprising a plurality of mapping relationships of previously stored query and corresponding guidance information, including selecting at least some of the plurality of models based at least in part on the similarities and using the selected models as index to search the database; and sending the obtained guidance information to be displayed to a user.