Diagnosis of Neuro-Degenerative Diseases Using Probabilistic Neural Network

Abstract

In recent years researchers have been given attention to present a non-invasive approach to deal with the diseases quickly and all the more unequivocally. The major cause of death of patients is due to the wrong diagnosis. Thus a right method is to be identified which would help the physicians to concentrate on the cause of illness and its diagnosis so as to abstain from squandering valuable time- that may be precious for the patient- on diagnosis. In this particular thesis, we have tried to build up a computerized methodology to deal with patients having problems in walking by analysing their gait signal. We chose four groups of patients, namely patients suffering from neuro-degenerative diseases such as Huntington’s disease, Parkinson’s disease and Amyotrophic Lateral Sclerosis and a group of healthy control subjects. So we have utilized Probabilistic Neural Network (PNN) as a classifier, to identify or differentiate the different patients and define the type of disease that they are suffering from, with accuracy in the range of 85- 95% so as to diagnose the diseases correctly and as a result, minimize the death rates