Parkinsons Disease Classification using Deep Learning

In the realm of neurological dysfunctional ailments, Parkinson’s disease (PD) affects patients, mostly old. The patients suffer from this disease that affects their cognition, behaviour and mobility as well. In most cases, patients show symptoms like tremors, rigidity, bradykinesia, and flat facial expression. This disease has affected a lot of people around the world, causing severe pain and trauma to patients and their families. Globally, this dreaded disease has affected millions of people in various advanced as well as poorer countries causing large scale loss of productive life and health among men and women, especially the middle-aged and the elderly. Looking at the disease profile and its aggressive tendencies to progress, early stage detection offers the best possible treatment options and possibly positive outcomes, helping patients secure a good quality of life. The present paper proposes a\ Convolutional neural network -based methodology to predict the severity of disease in patients suffering from Parkinson's by analyzing Telemonitoring Voice Data Set of patients sourced from UCI. In order to predict disease severity among patients, ‘Tensor Flow’ deep learning library of python has been used for implementing our neural network. Significantly, our method produces better accuracy values than the accuracy values found in earlier research works.