A three-way decision method with tolerance dominance relations in decision information systems

Area of Science:

• Decision Sciences
• Medical Informatics
• Artificial Intelligence

Background:

• Multi-attribute decision making (MADM) is increasingly applied to medical diagnosis.
• Three-way decision (3WD) theory offers a novel framework for MADM in healthcare by analyzing loss functions and decision matrices.
• Existing 3WD models often classify objects into only two states.

Purpose of the Study:

• To develop a three-way decision method incorporating tolerance dominance relations within ordered decision information systems.
• To enhance patient classification accuracy by considering both loss and gain functions, leading to net profit functions.
• To apply the proposed method to a real-world medical diagnosis problem, specifically Lymphography.

Main Methods:

• Utilized decision attributes to establish two states in ordered decision information systems.
• Developed net profit functions by integrating loss and gain considerations for each object.
• Formulated a class of three-way decisions based on tolerance dominance relations and MADM rules.

Main Results:

• A novel multi-attribute decision-making method was constructed using tolerance dominance relations.
• The proposed method was successfully applied to a Lymphography medical diagnosis case.
• Experimental evaluation demonstrated the feasibility and effectiveness of the new methodology.

Conclusions:

• The developed three-way decision method with tolerance dominance relations improves medical diagnosis accuracy.
• Integrating loss and gain functions provides a more comprehensive approach to patient classification.
• The methodology shows promise for practical applications in medical decision support systems.