Depression Diagnosis Using Optimization of Nonlinear EEG Features Based on Parametric Learning Tactics

Major Depressive Disorder (MDD) presents substantial challenges in mental health due to its intricate and multifaceted nature. In this study, we propose a framework for diagnosing MDD using the optimization of nonlinear features extracted from electroencephalogram (EEG) signals, utilizing a parametric learning tactic. Recognizing the inherent chaotic and nonlinear properties of EEG signals, we implemented a method involving dual windowing tactics and temporal overlap for feature extraction. This methodology enabled the identification of significant nonlinear features that effectively distinguish between MDD patients and healthy controls. Using the MUMTAZ dataset, we conducted rigorous feature selection and classification, employing K-Nearest Neighbor (KNN) and Support Vector Machine (SVM) classifiers. Our results indicate that the KNN classifier achieved the highest accuracy of 98.28%, demonstrating the efficacy of parametric learning tactics in optimizing nonlinear EEG feature extraction for accurate MDD diagnosis.