Word segmentation processing: a way to exponentially extend medical dictionaries

Area of Science:

• Medical Informatics
• Natural Language Processing (NLP)

Background:

• Medical lexicons are challenging to maintain due to the large number of compound words and new term creation.
• Existing medical dictionaries have two parts: grapheme-level identification and conceptual knowledge.
• Understanding medical expressions requires both identification and conceptual knowledge.

Purpose of the Study:

• To address the challenge of medical lexicon size in NLP.
• To develop a pragmatic approach for rapidly expanding the lexico-semantic component of medical dictionaries.
• To demonstrate the feasibility of a semi-automatic tool for medical term translation.

Main Methods:

• Development of a semi-automatic tool as a prototype.
• The tool focuses on translating medical diagnoses from French to ICD-9CM codes.
• Leveraging existing NLP techniques for lexico-semantic expansion.

Main Results:

• Successful creation of a prototype semi-automatic translation tool.
• Demonstrated feasibility of the proposed approach for lexicon expansion.
• The tool can translate French diagnoses into the ICD-9CM coding scheme.

Conclusions:

• The developed semi-automatic tool offers a pragmatic solution to expand medical lexicons.
• This approach can significantly improve the efficiency of medical NLP tasks.
• Facilitates better understanding and coding of medical diagnoses.