The ForEx++ based decision tree ensemble approach for robust detection of Parkinson’s disease

Abstract The progressive reduction of dopaminergic neurons in the human brain, especially at substantia nigra is one principal causes Parkinson’s Disease (PD). Voice alteration earliest symptoms found PD patients. Therefore, impaired subjects’ acoustic voice signal plays a crucial role detecting presence Parkinson's. This manuscript presents four distinct decision tree ensemble methods detection on trailblazing ForEx++ rule-based framework. Systematically Developed Forest (SysFor) and Penalizing Attributes Decision (ForestPA) approaches has been used for detection. proposed schemes efficiently identify positive subjects using primary features, viz . , baseline, vocal fold, time–frequency. A novel feature selection scheme termed Feature Ranking to Selection (FRFS) also combine filter wrapper strategies. FRFS encompasses Gel’s normality test rank selects outstanding features from time–frequency, fold groups. SysFor ForestPA forests underneath framework both ranking subset represents approaches, which expedite better overall impact segregating control subjects. It observed that forest ranked proved be robust scheme. models deliver highest accuracy 94.12% lowest mean absolute error 0.25, resulting an Area Under Curve (AUC) value 0.97.