A consultation system constructor for medical data analysis

Area of Science:

• Computer Science
• Medical Informatics
• Artificial Intelligence

Background:

• Expert systems are crucial for complex data analysis in specialized domains.
• Bridging the gap between expert knowledge and clinical decision-making requires robust modeling tools.
• Existing systems may lack flexibility in integrating diverse data analysis programs.

Purpose of the Study:

• To introduce the MAD system, designed to assist expert data analysts.
• To enable the creation of reasoning models for specific tasks in domains like epidemiology and image analysis.
• To facilitate the guidance of medical doctors in analyzing domain-specific data.

Main Methods:

• Expert knowledge is formalized at multiple levels, including domain descriptions and task-specific reasoning models.
• Reasoning models are implemented using propositional calculus rules.
• Meta-knowledge supports the knowledge acquisition process.
• The system integrates external data analysis programs seamlessly, requiring no user training for them.

Main Results:

• The MAD system effectively represents expert knowledge for building reasoning models.
• It supports the integration of external data analysis tools, enhancing usability for medical professionals.
• Demonstrated applications in epidemiology and image analysis showcase its practical utility.

Conclusions:

• MAD provides a flexible framework for expert data analysis and clinical decision support.
• The system's design facilitates knowledge acquisition and the application of formalized reasoning.
• Its ability to link with external software under MS-DOS makes it a versatile tool for specific analytical needs.