Item Graph Query Matching for Synonym-Aware Product Search
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current online concierge systems struggle to accurately identify items matching customer search queries due to mismatched search terms and limited information from warehouses, failing to account for synonyms and item relationships.
Innovation Solution
An online concierge system utilizes an item graph with nodes representing items and attributes, leveraging connections based on customer interactions to segment search queries, account for synonyms, and score candidate items for display, ensuring accurate item matching.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If the online concierge system uses limited information from the warehouse to identify items, then the system complexity is reduced, but the item identification accuracy deteriorates
Solution Approach 1:
The patent introduces an item graph as an intermediary data structure between the warehouse information and the search query processing. The item graph enriches limited warehouse information with additional attributes and relationships, enabling accurate item identification without increasing system complexity. The graph serves as a mediator that transforms sparse warehouse data into comprehensive item profiles.
Solution Approach 2:
The system performs preliminary actions by pre-building the item graph from available warehouse information before search queries are processed. This preliminary enrichment of item data with attributes and relationships allows the system to handle limited warehouse information effectively, improving identification accuracy without adding complexity during query processing.
2Ease of manufacture
If the online concierge system uses exact term matching from the warehouse, then the information processing is simplified, but the search coverage deteriorates
Solution Approach 1:
The patent changes the matching parameter from exact term matching to attribute-based matching. Instead of requiring precise term matches, the system matches search queries against item attributes in the graph, allowing synonyms and related terms to be recognized. This parameter change maintains processing simplicity while dramatically improving search coverage and adaptability.
Solution Approach 2:
The system adds another dimension to the matching process by introducing attribute relationships in the item graph. Instead of one-dimensional exact term matching, the system uses multi-dimensional attribute matching including synonyms, categories, and relationships, enabling comprehensive search coverage while keeping the matching mechanism manageable.
3Speed
If the online concierge system does not account for synonyms, then the processing speed is improved, but the item matching accuracy deteriorates
Solution Approach 1:
The system performs preliminary action by pre-processing and storing synonym relationships in the item graph during graph construction. When search queries are processed, the system leverages these pre-established synonym connections in the graph, achieving accurate item matching without the computational overhead of real-time synonym processing, thus maintaining high processing speed.
Data Source
AI summary
An online concierge system generates a graph connecting items with attributes of the items and other items. Hence, the graph includes nodes corresponding to attributes and nodes corresponding to items, with an item connected to attributes of the item in the graph. Example attributes include a brand, a category, a department, or any other suitable information about the item. When the online concierge system receives a search query to identify one or more items from a customer, the online concierge system parses the search query into combinations of terms and compares different combinations of terms to the graph to determine connections between different combinations of terms in the graph. Based on measures of connectedness between combinations of terms and connections in the graph, items are identified from one or more combinations of terms. Information about the identified items is presented to the customer.