Prediction of Parkinson Disorder Based on Kinect Arm Movement by Using Machine Learning Techniques

Abstract

Parkinson Disorder is a gradually increasing neurodegenerative disease. In this, motor and non-motor symptoms can impact on human basic functions to a wide variety of range. This article describes the required features and predicts the Parkinson Disorder in a patient with higher efficacy. In this work, mean vector velocity (MVV) was performed to analyze the essential features and crossed verified it by the Rank of Correlation. Previously many scientists analyzed Parkinson disorder using the fundamental classification models and also the different techniques to identify the postural tremor in the arm. It was found that they concentrated on the slow movement of the patients. But it has been proven that some abnormalities are also present in the medium movement. In this work the main tasks involve are (I) detecting four arm-features (II) calculating the mean instantaneous velocity of that four features and (III) performance evaluation of the four classification models under Artificial Neural Network. Finally we can able to predict the Parkinson disorder for patients with having good accuracy and sensitivity.