Document Subject Identification Using Peer Anchor Text Patterns
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for identifying the subject of a document are inadequate, as they rely on document titles, which may be absent, multiple, or misleading, and require human or computational processing of extensive content, making them impractical for large-scale factual information extraction from web-based documents.
Innovation Solution
A system and method that identifies a subject for a source document by analyzing a collection of peer documents from the same domain, extracting anchor text from linking documents, generating labels based on the longest-match anchor text, and applying patterns to the source document's title to determine its subject.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If document title is used as the subject identifier, then the process is simple and fast, but the accuracy is poor because titles may be absent, multiple, or misleading
Solution Approach 1:
The patent introduces peer documents from the same domain as an intermediary to mediate between the source document title and the final subject identification. These peer documents provide contextual information through their titles and anchor texts, helping to disambiguate misleading titles and identify the true subject when titles are absent or inaccurate
Solution Approach 2:
The system uses anchor texts from linking documents as feedback to validate and refine the subject identification. By analyzing what other documents reference the source document as, the system can confirm or correct the subject matter, creating a feedback loop that improves accuracy without significantly increasing processing time
2Measurement precision
If extensive content processing is performed to extract subject from document content, then the accuracy improves, but the complexity and processing time increase significantly
Solution Approach 1:
The patent extracts only the essential elements (titles and anchor texts) from peer documents rather than processing entire document contents. This selective extraction of relevant information maintains accuracy while dramatically reducing processing complexity and resource requirements
Solution Approach 2:
The system performs preliminary analysis by collecting peer document titles and anchor texts before attempting subject identification. This pre-processing step organizes relevant contextual information in advance, making the actual subject determination simpler and more accurate without requiring extensive content analysis
Data Source
AI summary
A system and method identifies a subject for a source document. The system and method identifies a collection of peer documents from the same domain as the source document. For each of the peer documents, a collection of linking documents containing a hyperlink to the peer document is identified. For each of the peer documents, a label is generated by choosing the longest-match anchor text of the linking documents. A pattern between the labels and the titles of the collection of peer documents is deduced. The subject of the source document is identified by applying the pattern to the title of the source document.