Symbolic Artificial Intelligence to Diagnose Tuberculosis Using Ontology.

Area of Science:

• Medical Informatics
• Artificial Intelligence in Medicine
• Infectious Disease Diagnostics

Background:

• Pulmonary Tuberculosis (PTB) is an infectious disease caused by Mycobacterium tuberculosis.
• Accurate PTB diagnosis is challenging due to symptom overlap with other lung conditions.
• Current diagnostic methods, relying solely on microbiological tests or X-rays, have limitations.

Purpose of the Study:

• To develop a Symbolic Artificial Intelligence (SAI) system for diagnosing PTB.
• To integrate diverse clinical and paraclinical data for improved diagnostic accuracy.
• To create a novel PTB ontology for comprehensive data storage and analysis.

Main Methods:

• Development of a PTB-specific domain ontology.
• Implementation of a knowledge base incorporating performance indicators.
• Utilizing a real-world database of over four years from Pondicherry hospital, India.
• Employing Symbolic Artificial Intelligence for diagnostic reasoning.

Main Results:

• The proposed SAI system demonstrates a potential for accurate PTB diagnosis.
• The system integrates clinical symptoms and paraclinical test results.
• The ontology facilitates the storage and retrieval of extensive patient data.

Conclusions:

• The developed SAI system offers a robust solution for diagnosing current and future PTB cases.
• Integrating comprehensive paraclinical data significantly improves diagnostic capabilities.
• This approach aids in identifying both PTB patients and other abnormal cases.