Diagnosing tuberculosis with a novel support vector machine-based artificial immune recognition system.

Area of Science:

• Medical diagnostics
• Artificial intelligence in healthcare
• Infectious disease research

Background:

• Tuberculosis (TB) is a leading infectious cause of death globally.
• Current TB diagnosis relies on slow culture-based methods, necessitating faster detection strategies.
• Artificial Immune Recognition System (AIRS) shows potential for disease diagnosis but requires accuracy improvements.

Purpose of the Study:

• To enhance the classification accuracy of AIRS for tuberculosis diagnosis.
• To introduce a novel hybrid system integrating support vector machines (SVM) with AIRS.

Main Methods:

• A hybrid AIRS-SVM model was developed for tuberculosis diagnosis.
• Patient data from the Pasteur laboratory of Iran (175 samples) was utilized.
• Performance was evaluated using 10-fold cross-validation, assessing accuracy, sensitivity, specificity, RMSE, Youden's Index, and AUC via WEKA software.

Main Results:

• The hybrid system achieved perfect classification with 100% accuracy, sensitivity, and specificity.
• Root Mean Squared Error (RMSE) was 0, indicating minimal error.
• Youden's Index and Area Under the Curve (AUC) were both 1, signifying optimal performance.

Conclusions:

• The developed hybrid AIRS-SVM model demonstrates exceptional efficacy in diagnosing tuberculosis.
• Achieving 100% sensitivity and specificity, this model can significantly aid medical professionals.
• This system offers a rapid and accurate diagnostic tool, crucial for effective tuberculosis control.