Diagnosis of Parkinson's disease using linear and non-linearfeatures and machine learning

Parkinson's disease (PD) is a progressive neurodegenerative disorder characterized bymotor symptoms such as tremors, rigidity, bradykinesia, and postural instability, as wellas non-motor symptoms including cognitive decline, sleep disturbances, and autonomicdysfunction [۱-۳]. Early diagnosis is critical for managing symptoms and slowing diseaseprogression, yet it remains a challenge due to the subtlety and variability of early-stagesymptoms [۴-۵]. Traditional diagnostic methods, relying heavily on clinical assessmentand neurological examination, often fall short in detecting the disease at its incipientstages. Consequently, there is a pressing need for more accurate and sensitive diagnostictools.