Search Result Diversification Using Category Similarity Matrices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing search technologies often fail to provide appropriately categorized and diverse search results, leading to unnecessary consumption of computing resources and repetitive user queries, which increases network latency and storage device I/O, especially when users are seeking items within similar categories.
Innovation Solution
A search engine leveraging similarity sensitive diversity to measure variation in item distribution across categories, using a cosine similarity matrix to improve search result distribution and reduce repetitive queries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If keyword search is used to return relevant items, then search queries can be processed, but search results may not be appropriately distributed across categories when items have similar attributes
Solution Approach 1:
The patent changes the parameter of category representation from simple keyword matching to cosine similarity calculations between category vectors. This transformation allows the system to measure the angular distance between categories in a multi-dimensional space, enabling more nuanced distribution of search results across semantically similar categories while maintaining search accuracy.
Solution Approach 2:
The patent introduces a new dimension by computing pairwise similarity matrices between categories using cosine similarity. This creates an additional layer of categorical relationships beyond traditional keyword matching, allowing the system to understand and distribute results across categories based on their vector-space relationships rather than just exact matches.
2Quantity of substance
If search systems return all relevant items, then completeness is achieved, but computing resources are unnecessarily consumed and network latency increases
Solution Approach 1:
The patent applies local quality by using similarity-sensitive diversity to identify and prioritize specific regions or segments of search results based on category diversity. Rather than uniformly processing all results, the system locally optimizes the presentation and ranking of items within different category groups, reducing the need to process and transmit every single result while maintaining completeness of relevant information.
3Reliability
If users submit multiple queries to find items in similar categories, then search completeness improves, but network latency and storage device I/O increase
Solution Approach 1:
The patent performs preliminary action by pre-computing and storing the pairwise similarity matrix between categories. This pre-processing step creates a lookup structure that allows the system to quickly determine category relationships and guide single queries to return comprehensive results across similar categories, eliminating the need for users to submit multiple sequential queries and thereby reducing network latency and storage I/O.
Data Source
AI summary
Similarity sensitive diversity is utilized to measure variation in a distribution of item listings along one or more categories. A similarity between category vectors of each category pair in a set of categories is determined and utilized to generate a pairwise similarity matrix. The pairwise similarity matrix may be pruned to remove category pairs below a threshold. Utilizing the pairwise similarity matrix, similarity sensitive diversity between one or more items of a plurality of items may be determined. In various aspects, the similarity sensitive diversity may be utilized to: generate a list of relevant items in an appropriate distribution, suggest refinements of a search query; generate navigation modules; categorize or recategorize the plurality of items; or generate autosuggestions.