Late feature fusion using neural network with voting classifier for Parkinson's disease detection

Area of Science:

• Neurology
• Artificial Intelligence
• Machine Learning

Background:

• Parkinson's disease (PD) is a progressive neurological disorder characterized by cell death in the midbrain.
• Early detection of PD is crucial for managing disease progression and improving patient outcomes.
• Artificial intelligence (AI) offers potential for analyzing diverse datasets (text, speech, image) for disease classification.

Purpose of the Study:

• To develop and validate a hybrid AI model for accurate early detection and classification of Parkinson's disease.
• To enhance diagnostic accuracy by integrating deep learning and machine learning techniques.
• To explore the efficacy of combining handwriting analysis and motor symptom data for PD identification.

Main Methods:

• Feature extraction using Convolutional Neural Networks (CNN) and attention mechanisms.
• Processing of motion signals with a combined CNN and Long Short-Term Memory (LSTM) model.
• Classification using ensemble methods including Random Forest, Logistic Regression, Support Vector Machine, Extreme Gradient Boosting, and a voting classifier.

Main Results:

• The proposed hybrid model achieved exceptional performance on PD handwriting and motion datasets.
• Achieved 99.95% accuracy, 99.99% precision, 99.98% sensitivity, and 99.95% F1-score.
• Demonstrated the effectiveness of the integrated feature extraction and voting classifier approach.

Conclusions:

• The methodology integrating feature extraction from handwriting and motor symptoms, followed by a voting classifier, is highly effective for Parkinson's disease classification.
• This AI-driven approach shows significant promise for accurate and early diagnosis of PD.
• The study highlights the potential of advanced AI techniques in neurological disease detection.