Document Ranking via Term Relationship Graphs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current search systems struggle to provide diverse search results for ambiguous queries and fail to effectively reorder search results to introduce a broader range of information, while also neglecting relationships between documents without hyperlinks.
Innovation Solution
The method involves identifying local and global term relationships within documents, generating graphs to represent these relationships, and determining scores for documents based on these relationships, allowing for the reordering of search results to present a more diverse range of information and highlighting related terms as navigational references.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If search results are ranked based on traditional relevance factors, then the search engine can efficiently return matching web pages, but the search results lack diversity for ambiguous queries
Solution Approach 1:
The patent segments the search result ranking process into multiple independent scoring components: traditional relevance scoring, term relationship scoring, and document relationship scoring. Each component evaluates different aspects of document quality and diversity, allowing the system to maintain high efficiency while improving diversity through the combination of these segmented scoring mechanisms.
Solution Approach 2:
The patent introduces a new dimension to search result evaluation by incorporating term relationship graphs and document relationship graphs. These graphs add structural dimensions to the traditional flat relevance scoring, enabling the system to evaluate documents based on their contextual relationships and information diversity, thereby resolving the contradiction between efficiency and diversity.
2Adaptability or versatility
If search results are reordered to present diverse information, then the range of information presented is broader, but the original relevance ranking may be compromised
Solution Approach 1:
The patent merges multiple scoring mechanisms into a unified ranking system. The final document score combines traditional relevance scores with relationship-based scores from term and document graphs. This merging allows the system to maintain ranking accuracy by preserving the original relevance foundation while enhancing information diversity through the additional relationship-based scoring dimensions.
Solution Approach 2:
The patent changes the parameters used for document evaluation by incorporating relationship-based metrics alongside traditional relevance parameters. By adjusting and combining multiple scoring parameters (relevance, term relationship strength, document relationship strength), the system achieves both accurate ranking and diverse information presentation.
3Ease of operation
If relationships between documents without hyperlinks are determined, then navigational references are improved, but the system complexity increases
Solution Approach 1:
The patent introduces term relationship graphs as an intermediary structure to infer document relationships without requiring direct hyperlinks. These graphs serve as a mediator that captures semantic relationships between terms across documents, enabling the system to establish document connections through shared terminology and contextual relationships, thereby improving navigation without proportionally increasing system complexity.
Solution Approach 2:
The system performs self-service by automatically constructing term relationship graphs and document relationship graphs from the document corpus without manual intervention. The relationship extraction process is automated through graph-based algorithms that infer connections based on term co-occurrence and document similarity, reducing the operational burden while improving navigational capabilities.
Data Source
AI summary
Methods, systems, and apparatus, including computer program products, for scoring documents. A plurality of documents with an initial ordering is received. Local term relationships between terms in the plurality of documents are identified, each local term relationship being a relationship between a pair of terms in a respective document. Relationships among the documents in the plurality of documents are determined based on the local term relationships and on the initial order of the documents. A respective score is determined for each document in the plurality of documents based on the document relationships.