Diagnosis of Parkinson’s Disease Using Deep Neural Network Model

Parkinson’s disease is a neuro-degenerative disorder that effects central nervous system and is observed in many people worldwide. PD diagnosis is complex for the clinicians as it requires meticulous analysis of the patient. Though there are many characteristics and symptoms that indicate the disease, voice characteristics play a major role among the predictive characteristics. Person with PD experiences several vocal degradations like shaky and low speech. Voice analysis offers the additional benefit of being non-invasive, low cost and simple to diagnose. Many enthusiastic and great researchers have created new models and improved existing models in this area, and there is a vast amount of research in this field all over the world. We created an optimized Deep neural network (which is referred as Opt-DNN in rest of the paper) model and compared it to various algorithms such as random forests, SVM, XG Boost, and KNN in this paper. Among all the algorithms used, the proposed model turned up to be the best algorithm with accuracy 95.14.