Dynamic Search String Generation for Retail Context
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional search systems in online retail are limited in retrieving relevant product reviews as they rely on exact search terms, failing to capture relevant information when different users describe product features using different words, making it burdensome for shoppers to find useful reviews.
Innovation Solution
A computerized method generates a dynamic search string based on a context-specific dictionary, matching user queries to relevant terms, allowing for context-relevant search results to be returned, even if the search query does not contain exact terms used in reviews.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If exact text search is used to find product reviews, then search precision is improved, but search recall deteriorates because different users describe the same product features using different words
Solution Approach 1:
The patent introduces a synonym dictionary as an intermediary between the user's search query and the product review database. The system expands the search query by replacing terms with their synonyms from the dictionary, enabling matches even when exact wording doesn't match. This mediator layer bridges the gap between different user vocabularies while maintaining search accuracy.
Solution Approach 2:
The system dynamically changes the search parameters by generating multiple search queries with different term variations. Instead of using a single exact search term, the system transforms the original query into multiple expanded queries using synonym substitution, thereby searching with varied parameters to capture more comprehensive results.
2Device complexity
If simple text search is used, then device complexity is reduced, but ease of operation deteriorates because shoppers must know the exact terms used in reviews
Solution Approach 1:
The system performs self-service by automatically expanding search queries using the synonym dictionary without requiring user intervention. The shopping system itself manages the complexity of term variations, making the search process easier for users while maintaining low operational complexity through automated dictionary-based expansion.
3Loss of information
If context-based dynamic search string generation is implemented, then search recall is improved, but device complexity increases due to dictionary matching and term expansion processes
Solution Approach 1:
The synonym dictionary is prepared and stored in advance, containing pre-defined term relationships. This preliminary action allows the system to quickly perform lookups during search operations without complex real-time analysis, reducing the computational complexity while maintaining high search recall through pre-computed synonym mappings.
Data Source
AI summary
A method of generating search results, comprising; receiving, by a computing system, an instruction to generate search results relating to a SKU stored in a database; extracting, by the computing system, a context based on information relating to the SKU; generating, by the computing system, a dynamic search string, by: generating a key based on the instruction; identifying a dictionary based on the context, the dictionary containing a plurality of dictionary term; and generating the dynamic search string by determining the dictionary term corresponding to the key; searching, by the computing system, using the dynamic search string, in content data; returning results from the content data matching the dynamic search string; and generating for display on a client device, the results.