Performance evaluation of machine learning algorithms for Parkinson’s disease identification model

Abstract

Though Medical sector is referred to as one of the most advance sector in the Country, there are some implications which shows disorientation of medical sector and early prediction of various ailments. There are some diseases which does not have any preventive measures and are incurable till date. Parkinson’s disease is one of those disease that is incurable till date and its clinical diagnosis requires early detection of symptoms which may vary according to age parameter. Parkinson’s disease is one of the incurable diseases which affects the Central Nervous System (CNS) of Human brain which leads to various symptoms like barely noticeable tremor on limbs, slowness in movement, improper balance and posture of body, Loss of automated movement and sometimes change in vocal frequency. With theadvent of Machine Learning (ML) and Artificial Intelligence (AI) technique are used for diagnosis of brain diseaseswhich is helping the doctor for early detection of the diseases.Considering the performance evaluation parameters, we have given equal importance to every performance measures used. The dataset for the Parkinson’s disease detection is taken from Oxford UCI Machine repository. Analysis and Comparative study show that all the algorithms can detect the person is affected by the Parkinson’s disease or not. Out of them random forest algorithm yielded highest performance measures.