Parkinson’s Disease Prediction Using Machine Learning

Abstract

A neurological condition of the brain is Parkinson's disease. It causes the body to shake, the hands to shake, and it makes the body stiff. At this advanced level, there is still no viable treatment or cure. Only when treatment is initiated at the earliest stage of the disease is it effective. These could potentially save a life in addition to lowering the cost of the illness. The majority of ways can identify Parkinson's disease after it has advanced, which results in basal ganglia, which regulates movement of the body with a small quantity of dopamine, losing about 60% of their dopamine. The presence of diminutive precursors to chronic diseases, especially neurologically based ailments such as Parkinson’s can indicate and diagnose them in their earliest stages. Parkinson’s disease (PD) is characterized by symptoms such as spasms in the limbs, jaw, and head, rigidity in the limbs and trunk, slow movement, etc. It is important to notice these preliminary symptoms early on to avoid developing Parkinson's disease. This project proposes algorithms used on the dataset to predict PD. Two kinds of datasets are used; one voice dataset and another spiral drawing dataset, and algorithms are used in these datasets to predict the disease and to show the results a user-friendly Web-Application is developed. The algorithms used are K-Nearest Neighbours (KNN) on voice dataset implementation and Random Forest on spiral drawing implementation.