Related Experiment Videos
Evaluation of a Meta-1-based automatic indexing method for medical documents
1Section of Medical Informatics, University of Pittsburgh School of Medicine, PA 15261.
Computers and Biomedical Research, an International Journal
|August 1, 1992
Summary
MetaIndex is an automatic indexing program that enhances document retrieval by creating symbolic representations using the Unified Medical Language System Metathesaurus. This approach aims to improve how documents are represented for better search performance.
Area of Science:
- Medical Informatics
- Information Retrieval
- Natural Language Processing
Background:
- Automatic indexing is crucial for efficient document retrieval in large medical collections.
- Existing methods may not fully capture the semantic content of medical documents.
- The Unified Medical Language System Metathesaurus (Meta-1) offers a rich source of medical concepts.
Purpose of the Study:
- To introduce MetaIndex, an automatic indexing program designed to improve document retrieval.
- To utilize a symbolic representation of documents based on Meta-1 concepts.
- To enhance document representation for better retrieval performance.
Main Methods:
- Developed MetaIndex, an automatic indexing program.
- Employed a transition network parser to recognize a language derived from Meta-1 concepts.
- Utilized a hierarchy of medical concepts from Meta-1 for document content representation.
Main Results:
- MetaIndex successfully creates symbolic representations of documents.
- The program leverages Meta-1 for concept-based indexing.
- Evaluation of MetaIndex on the Slice of Life medical image collection was performed.
Conclusions:
- MetaIndex offers a novel approach to automatic indexing for medical documents.
- Symbolic representation using Meta-1 concepts shows promise for improving document retrieval.
- Further evaluation is warranted to confirm performance benefits.