Simplified representation of concepts and relations on screen

Area of Science:

• Medical Informatics
• Natural Language Processing
• Knowledge Representation

Background:

• Automated diagnostic code generation necessitates interpreting noun phrases and representing word sense content.
• Existing knowledge-based systems often prioritize calculus over data structures, complicating the systematic modeling of complex natural language information.

Purpose of the Study:

• To present a novel notation system for representing complex data structures and multi-branched rules for automated diagnostic coding.
• To explain the principles of the "concept molecule" (CM) notation and compare it with conceptual graphs.

Main Methods:

• Development of a notation system using "concept particles" and later evolving to "concept molecules" (CM).
• Analysis and representation of word sense content from noun phrases for systematic data modeling.
• Comparative analysis of the CM notation with conceptual graphs (CGs).

Main Results:

• The "concept molecule" (CM) notation allows for the simple denotation of complex data structures and multi-branched rules.
• CM notation emphasizes data structures for systematic modeling, simplifying the coding process compared to complex information incorporation.
• The study provides a detailed explanation of CM principles and contrasts them with CGs.

Conclusions:

• The "concept molecule" (CM) notation offers an effective method for representing complex information required for automated diagnostic coding.
• CM notation provides a systematic approach to data modeling, overcoming limitations of earlier methods and simplifying rule representation.
• This work contributes a robust knowledge representation technique for advancing automated diagnostic systems.