Detection of Parkinson's Disease from EEG Signals with EEMD Using Machine Learning and Deep Learning Techniques

Parkinson's Disease (PD) is a progressive neurodegenerative disorder characterized by tremors, stiffness, and difficulty with movement, impacting mil-lions worldwide. Electroencephalography (EEG) signals can help detect Parkinson's disease by revealing abnormal brain wave patterns associated with motor dysfunction. This abstract examines the use of EEG data for PD classification, giving attention to the analysis of frequency bands using machine learning (ML) and deep learning (DL) methods. This research suggests a novel ML and DL models based on Extended Empirical Mode Decomposition (EEMD) signal processing method to classify PD in its early stages. The aim of the proposed method is to divide patients into two groups: PD positive and negative. The developed model uses a Machine learning algorithms including Random Forest, Gradient Boost, SVM and deep learning algorithms including CNN, DNN developed for the detection of PD among patients. Random Forest and Gradient Boost excelin ensemble learning. Enriching the performance of the modela decomposition method - Extended Empirical Mode Decomposition (EEMD), is employed to analyze complex EEG signals. EEMD's application enhances feature extraction for ML and DL algorithms, facilitating accurate PD detection. The diagnostic accuracy is further improved using CNN and DNN models, which identify complex patterns in EEG data. This integrated approach showcases promise for early classification and continuous monitoring, contributing to advancements in understanding and managing PD. The proposed CNN model achieved a greater of 0.98 accuracy metric.