Entity Name Recognition via Pattern Matching and Whitelist Validation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for recognizing entity names in documents are inadequate due to the impracticality of human editing, under-inclusiveness, and over-inclusiveness of relying on single sources like the Internet Movie Database, which fail to efficiently organize and analyze factual information across various domains.
Innovation Solution
A method and system that select documents sharing a common pattern in their titles, determine name candidates by applying this pattern, and match them with a collection of entity names (white list) to validate and add valid candidates to the list, enhancing the recognition of entity names across multiple domains.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If human editors are used to review documents for entity name recognition, then accuracy of entity name identification is improved, but productivity and scalability deteriorate due to the vast volume of documents
Solution Approach 1:
The patent segments the entity name recognition task into multiple processing stages: (1) extracting candidate entity names from document titles using pattern matching, (2) filtering candidates through multiple criteria including whitelist matching, (3) validating candidates against multiple data sources. This segmentation allows automated processing while maintaining accuracy through layered verification.
Solution Approach 2:
The patent introduces an intermediary validation layer that mediates between automated extraction and final entity recognition. The system uses multiple data sources (whitelists, knowledge bases) as intermediaries to verify extracted candidates, combining automated efficiency with human-curated accuracy without requiring human editors to review every document.
2Reliability
If entity names are extracted from a single reputable source like the Internet Movie Database, then reliability of extracted entity names is improved, but adaptability and coverage of different entity types deteriorate
Solution Approach 1:
The patent creates a universal entity recognition system that handles multiple entity types (people, places, organizations, movies, etc.) through a single multi-source validation framework. The system uses multiple whitelists and data sources, each specialized for different entity types, combined in a unified processing pipeline that adapts to various domains while maintaining reliability through cross-validation.
Solution Approach 2:
The patent combines multiple data sources and validation methods into a composite recognition system. Instead of relying on a single source, the system integrates multiple whitelists, knowledge bases, and validation criteria, creating a composite approach that leverages the strengths of each source while compensating for individual limitations through mutual verification.
3Adaptability or versatility
If entity names are extracted from multiple sources to increase coverage, then adaptability is improved, but measurement precision deteriorates due to inclusion of non-entity names
Solution Approach 1:
The patent performs preliminary filtering and validation of candidate entity names before final recognition. The system pre-processes extracted candidates by checking against multiple whitelists and validation criteria, eliminating non-entity names early in the process. This preliminary action ensures that only high-quality candidates proceed to final recognition, maintaining precision while covering diverse entity types.
Solution Approach 2:
The patent implements feedback mechanisms where validation results from multiple sources feed back into the recognition process. The system uses whitelist matching, knowledge base verification, and cross-source validation to provide feedback on candidate quality, continuously refining the recognition process and improving precision through iterative verification against multiple authoritative sources.
Data Source
AI summary
A system, method, and computer program product for recognizing entity names from a plurality of documents. Embodiments of the methods comprise selecting a selection of documents from a plurality of documents, the selection of documents sharing a common pattern in their titles. The method further comprises determining a name candidate for each document in the selection by applying the common pattern to the title of the document, and matching the name candidates with a collection of entity names (the white list). Responsive to determining a match between the name candidates and the entity names in the white list, the method determines that the name candidates are valid entity names. In one embodiment, the name candidates are added to the white list after being determined to be valid entity names.