Hybrid Disease Diagnosis Using Multiobjective Optimization with Evolutionary Parameter Optimization.

Area of Science:

• * Artificial Intelligence in Healthcare
• * Machine Learning for Medical Diagnosis
• * Computational Intelligence in Medicine

Background:

• * Growing adoption of e-Healthcare and telemedicine necessitates accurate, intelligent disease diagnosis systems.
• * Limitations of individual machine learning classifiers for comprehensive disease classification are widely acknowledged.
• * Ensemble classification techniques are emerging as a promising approach to overcome individual model weaknesses.

Purpose of the Study:

• * To propose and evaluate novel hybrid intelligent systems (HISs) for disease diagnosis.
• * To optimize parameters of Support Vector Machine (SVM) and Multilayer Perceptron (MLP) classifiers using evolutionary algorithms.
• * To assess the performance of HISs based on prediction accuracy, sensitivity, and specificity.

Main Methods:

• * Development of six hybrid disease diagnosis systems by optimizing SVM and MLP parameters with three evolutionary algorithms.
• * Utilizing prediction accuracy, sensitivity, and specificity as key performance metrics.
• * Evaluation of the proposed HISs on 11 benchmark medical datasets.

Main Results:

• * The proposed hybrid intelligent systems consistently outperformed existing methods across multiple datasets.
• * Demonstrated significant improvements in disease prediction accuracy, sensitivity, and specificity.
• * Statistical tests confirmed the efficacy and superiority of the developed HISs.

Conclusions:

• * Hybrid intelligent systems integrating optimized SVM and MLP classifiers offer enhanced disease diagnosis capabilities.
• * The proposed approach effectively addresses the limitations of single machine learning models in complex medical scenarios.
• * These findings support the potential of advanced machine learning ensembles in improving e-Healthcare diagnostic accuracy.