Metaheuristics with Deep Learning-Enabled Parkinson's Disease Diagnosis and Classification Model.

Area of Science:

• Neurology
• Artificial Intelligence
• Biomedical Engineering

Background:

• Parkinson's disease (PD) significantly impacts motor functions like writing, speech, and muscle stiffness.
• Early detection of PD is crucial for improving patient quality of life and managing disease progression.
• Existing diagnostic methods often rely on voice, speech, and writing exams, with a need for enhanced accuracy.

Purpose of the Study:

• To present an improved sailfish optimization with deep learning (ISFO-DL) model for accurate PD diagnosis and classification.
• To enhance the early identification of Parkinson's disease for improved patient outcomes.
• To leverage metaheuristic algorithms and deep learning for robust PD detection.

Main Methods:

• Developed an improved sailfish optimization (ISFO) algorithm to identify optimal feature subsets for classification.
• Employed a deep learning (DL) model, specifically a rat swarm optimizer (RSO) with bidirectional gated recurrent unit (BiGRU), as the classifier.
• Validated the ISFO-DL model using a benchmark Parkinson's dataset.

Main Results:

• The ISFO algorithm effectively derived an optimal feature subset, maximizing classification accuracy.
• The ISFO-DL model demonstrated enhanced classification performance in identifying Parkinson's disease.
• Experimental results confirmed the efficacy of the proposed model on a benchmark dataset.

Conclusions:

• The proposed ISFO-DL model offers a promising approach for the earlier and more accurate identification of Parkinson's disease.
• This technique can significantly contribute to improving survival rates and patient management for individuals with PD.
• The integration of metaheuristic optimization and deep learning provides a powerful tool for neurological disorder diagnosis.