Pneumothorax prediction using a foraging and hunting based ant colony optimizer assisted support vector machine.

Area of Science:

• Medical diagnostics
• Artificial intelligence
• Computational biology

Background:

• Percutaneous needle lung biopsy (PNLB) is invasive and carries risks, with pneumothorax being a common complication.
• Accurate and timely diagnosis of pneumothorax is crucial for patient outcomes.
• Existing diagnostic methods may benefit from enhanced computational approaches for improved accuracy.

Purpose of the Study:

• To develop an advanced optimization algorithm for improved feature selection in medical diagnostics.
• To propose a novel hybrid optimization algorithm, SCACO, integrating slime mould foraging and collaborative hunting strategies.
• To introduce a Support Vector Machine (SVM) classifier, bSCACO-SVM, for pneumothorax diagnostic prediction.

Main Methods:

• Development of the Slime Mould foraging and Collaborative Ant Colony Optimization (SCACO) algorithm.
• Creation of a binary version of SCACO (bSCACO) for feature selection.
• Integration of bSCACO with an SVM classifier for pneumothorax prediction (bSCACO-SVM).
• Performance evaluation of SCACO against nine basic and nine variant algorithms.

Main Results:

• SCACO demonstrated improved convergence accuracy and solution quality compared to standard Ant Colony Optimization (ACO).
• The adaptive collaborative hunting strategy enhanced ACO's ability to escape local optima.
• bSCACO-SVM exhibited robust classification prediction capacity on public datasets.
• The proposed method showed successful application in tuberculous pleural effusion diagnostic prediction.

Conclusions:

• The novel SCACO algorithm offers enhanced performance for optimization tasks.
• bSCACO-SVM provides a reliable tool for pneumothorax diagnostic prediction.
• This approach holds potential for improving the diagnosis of lung conditions like tuberculous pleural effusion.