Search Diversity Index for Query Expansion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current information retrieval systems, such as online shopping and auction websites, face challenges in accurately matching user queries with relevant items due to differences in vocabulary used by buyers and sellers, leading to ineffective connections between buyers and sellers.
Innovation Solution
A search engine module with a diversity identification module measures the diversity of query terms and result sets to expand searches by suggesting alternative queries with related terms, using diversity indices like Simpson's and Shannon's indices to ensure sufficient diversity and improve search accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional search engines are used to match queries with items, then the search process is simple and fast, but the matching accuracy deteriorates due to vocabulary differences between buyers and sellers
Solution Approach 1:
The patent introduces an intermediary data model that acts as a bridge between buyer queries and seller item descriptions. This data model includes normalized attributes and a mapping layer that translates different vocabularies into a common representation, enabling accurate matching without requiring complex direct comparison between diverse query and item terminologies
Solution Approach 2:
The patent transforms the search approach by changing parameters from direct text matching to attribute-based matching. It converts unstructured item descriptions and queries into structured attribute-value pairs, allowing the system to match based on semantic equivalence rather than literal text overlap, thereby improving accuracy while managing complexity through standardization
2Productivity
If item descriptions are kept short and uniquely defined by sellers, then the listing process is efficient and easy to maintain, but the searchability deteriorates due to limited information and vocabulary diversity
Solution Approach 1:
The patent segments item descriptions into distinct attributes (e.g., category, specifications, features) that can be independently extracted and stored. This segmentation allows the system to maintain simple seller input while creating a structured data model that enables precise searching through attribute-based queries rather than requiring sellers to write comprehensive descriptive text
Solution Approach 2:
The patent creates a universal data model that serves multiple functions: it stores seller-provided information in a standardized format, enables diverse query types (exact match, partial match, attribute-based), and supports both simple and complex search operations. This multi-functional model allows short seller descriptions to be effectively utilized for various search scenarios without requiring sellers to provide extensive information
Data Source
AI summary
A method and system for a search based on diversity is described. The system calculates a diversity index for a result set comprising results for a query. The diversity index is a measure of diversity among the results of the query. The diversity index relates to differences among the results. The system compares the diversity index to a threshold value. Finally, the system expands the query with one or more query terms based on the comparison.