Medical Synonym Mining via Clustering and Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The complexity of natural language, particularly the phenomena of synonymy and polysemy, leads to challenges in automatically understanding and analyzing medical texts due to the lack of a unified word specification, resulting in multiple expressions for the same medical entities, which hinders the processing of medical data.
Innovation Solution
A synonym mining method that involves recognizing named entities, clustering them based on word form and context similarity, and filtering to obtain a synonym set, which can be used to construct a synonym dictionary for normalizing medical entities and improving natural language processing tasks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If synonym clustering is performed without filtering, then the quantity of synonym candidates increases, but the purity of the synonym set deteriorates
Solution Approach 1:
The patent segments the synonym mining process into distinct stages: initial broad clustering to generate candidate sets, followed by filtering processes that segment the candidate set into pure synonyms and non-synonyms. This segmentation allows the system to first maximize candidate quantity, then systematically purify the results through multiple filtering passes.
Solution Approach 2:
The patent introduces intermediary filtering mechanisms that act as mediators between the clustering process and the final synonym set. These filters include statistical significance tests, context similarity thresholds, and manual verification steps that intermediate the raw clustering output and the purified synonym set, thereby maintaining both quantity and purity.
2Manufacturing precision
If strict filtering criteria are applied to obtain pure synonym sets, then the purity of the synonym set improves, but the quantity of synonyms decreases
Solution Approach 1:
The patent performs preliminary actions by generating an extensive candidate synonym set before applying strict filtering criteria. The initial clustering phase creates a superset that includes all potential synonyms, ensuring that even strict subsequent filtering will not miss valid synonyms. This preliminary expansion compensates for the quantity loss during filtering.
Solution Approach 2:
The patent applies partial filtering by using multiple passes with varying strictness levels. Rather than applying one extremely strict filter that would eliminate too many candidates, the system uses successive filtering passes that progressively purify the set while retaining acceptable quantity. Some filters are applied partially to subsets of candidates based on preliminary assessments.
3Manufacturing precision
If multiple filtering processes are performed on synonym candidates, then the purity of the synonym set improves, but the complexity of the processing increases
Solution Approach 1:
The patent segments the filtering process into multiple independent, modular filtering steps rather than one complex monolithic filter. Each filter handles a specific aspect of purity verification (statistical significance, context similarity, lexical relationships), making the overall complex process manageable through segmentation into discrete, reusable components.
Solution Approach 2:
The patent performs preliminary assessments of candidate synonyms using quick, low-complexity filters before applying more complex filtering criteria. This staged approach applies simple filters first to eliminate obvious non-synonyms, then progressively applies more complex filters only to remaining candidates, reducing overall processing complexity while maintaining purity.
4Productivity
If automated clustering is used to group named entities, then the productivity of synonym mining improves, but the precision of synonym identification deteriorates
Solution Approach 1:
The patent introduces intermediary verification steps between automated clustering and final synonym identification. These intermediaries include statistical significance tests, context similarity measurements, and expert review processes that mediate between the automated clustering output and the final verified synonym sets, thereby recovering precision lost through automation.
Solution Approach 2:
The patent implements feedback loops where the results of precision verification are fed back into the clustering process. When automated clustering produces impure synonym sets, the filtering and verification results provide feedback that adjusts clustering parameters or triggers re-clustering, thereby improving precision while maintaining productivity through iterative refinement.
Data Source
AI summary
Disclosed are a synonym mining method, an application method of a synonym dictionary, a medical synonym mining method, an application method of a medical synonym dictionary, a synonym mining device and a storage medium. The synonym mining method includes: performing a recognition process on corpus data to obtain a named entity set of at least one category, wherein the named entity set of each category includes a plurality of named entities; performing a clustering process on the plurality of named entities in the named entity set of the each category to obtain a synonym candidate set corresponding to the each category; and performing, based on a word from similarity and a context similarity, a filtering process on the synonym candidate set corresponding to the each category to obtain a synonym set corresponding to the each category.