Stack Ensemble Oriented Parkinson Disease Prediction Using Machine Learning Approaches Utilizing GridSearchCV-Based...

Area of Science:

• Neurology
• Computer Science
• Artificial Intelligence

Background:

• The COVID-19 pandemic exacerbated mental health issues and highlighted the need for better awareness of neurological disorders like Parkinson's disease.
• Existing machine learning and deep learning models for Parkinson's disease prediction require optimization.
• Parkinson's disease is a significant nervous system disorder affecting the elderly population.

Purpose of the Study:

• To explore and compare various machine learning techniques for predicting Parkinson's disease onset.
• To develop and evaluate an optimized ensemble stacking model for improved prediction accuracy.
• To analyze the performance of deep learning models in detecting Parkinson's disease.

Main Methods:

• Implemented and compared traditional classifiers: naive Bayes, logistic regression, linear-support vector machine, kernelizing support vector machine, and multilayer perceptron.
• Developed an ensemble stacking model using hyper-tuned classifiers and extreme gradient boosting, optimized with GridSearchCV.
• Utilized a convolutional neural network for deep learning analysis on benchmark datasets (UCI Parkinson's, spiral, wave).

Main Results:

• The ensemble stacking approach achieved a high accuracy of 94.87%.
• Performance metrics including accuracy, precision, recall, and F1 score were used for evaluation.
• Confusion matrices and graphs were generated for visualization of model performance.

Conclusions:

• Ensemble stacking models demonstrate significant potential for enhancing Parkinson's disease prediction accuracy.
• Further research and optimization of machine learning and deep learning models are crucial for early detection and management of Parkinson's disease.
• The study provides a robust framework for utilizing advanced computational techniques in neurological disorder diagnostics.