PD Care: A Framework for Advance Screening of Parkinson’s Patient from Vocal Dysphonia

Parkinson’s Disease (PD) is a progressive neurological disorder that seeds dopamine loss in the brain and with decrement of dopamine, the traits of the disease become prominent. In 90% of cases, the voice impairment increases with the progression of the disease. There is no standardized test for early diagnosis of Parkinson’s Disease and conventional method of diagnosis encompasses observing daily actions, motor function, neurological factors to evaluate the progression of the disease. This study proposed a framework on applying classification model for early screening of Parkinson’s disease from the patient’s voice dataset and comparative analysis of applied algorithms to ascertain the best-performed algorithm. The dataset employed in this study provided by University of Oxford partnering with National Centre for Voice and Speech, Denver, Colorado. It is observed that K-Nearest Neighbors performed with 95.92% accuracy with 25% test data. In addition, the features such as MDVP: Jitter, RPDE, DFA, MDVP: Shimmer, contributes to higher accuracy, precision. We believe such frame-work and analysis will direct development of effective intelligent analytics-based system for early screening of Parkinson’s patient.