Diagnosis support using Fuzzy Cognitive Maps combined with Genetic Algorithms

Area of Science:

• Computational intelligence
• Medical informatics
• Speech pathology

Background:

• Existing medical diagnosis support systems face challenges with ambiguous decision-making.
• Fuzzy Cognitive Maps (FCMs) offer a framework for modeling complex systems but require enhancement for dynamic concept interaction.
• Genetic Algorithms (GAs) are powerful optimization tools applicable to complex problem-solving.

Purpose of the Study:

• To develop a novel hybrid modeling methodology for enhanced medical diagnosis decision support.
• To integrate Competitive Fuzzy Cognitive Maps (CFCMs) with Genetic Algorithms (GAs) for improved concept interaction.
• To create more dependable Advanced Medical Diagnosis Support Systems (AMDSS) capable of handling unclear decisions.

Main Methods:

• A hybrid approach combining CFCMs with GAs was developed.
• A new algorithm was proposed to achieve synergy between CFCMs and GAs.
• The methodology was applied to model and test a differential diagnosis problem in speech pathology.

Main Results:

• The hybrid methodology successfully modeled a differential diagnosis problem.
• The developed Advanced Medical Diagnosis Support Systems demonstrated enhanced dependability.
• The system effectively handled situations with non-distinct decisions in language impairment diagnosis.

Conclusions:

• The proposed hybrid modeling methodology offers a robust solution for complex medical diagnosis.
• Integrating CFCMs with GAs significantly improves the performance of diagnosis support systems.
• This approach shows promise for advancing the diagnosis of language impairments in speech pathology.