Text Mining the History of Medicine

Area of Science:

• Digital Humanities
• Medical Informatics
• Computational Linguistics

Background:

• Large-scale digitization of historical text archives provides rich data, but efficient exploration remains challenging.
• Text mining (TM) offers automated semantic information extraction (concepts, relationships) for enhanced search functionality.
• Applying TM to historical medical texts presents unique challenges due to linguistic evolution.

Observation:

• Historical medical texts exhibit evolving vocabulary, terminology, and linguistic structures compared to modern texts.
• Developing resources that account for concept variations and relationships across different historical periods is crucial.
• A modular processing pipeline is needed for robust semantic information detection in diverse historical documents.

Findings:

• Novel resources were created using diverse historical medical documents to capture evolving concept expressions and relationships.
• A modular TM processing pipeline was developed for robust semantic analysis of historical medical texts.
• The pipeline was applied to large medical archives, forming the basis for a semantically-oriented search system.

Implications:

• The developed resources and pipeline facilitate the semantic analysis of historical medical literature.
• A publicly accessible, semantically-oriented search system enhances research accessibility to historical medical data.
• The processing pipeline and resources are available for broader research and integration into platforms like Argo TM.