SVM based machine learning approach to identify Parkinson's disease using gait analysis

Parkinson's Disease (PD) is a neuro-degenerative disease which affects a persons mobility. Tremors, rigidity of the muscles and imprecise gait movements are characteristics of this disease. Past attempts have been made to classify Parkinsons disease from healthy subjects but in this work, effort was made to focus on the specific gait characteristics which would help differentiate Parkinsons Disease from other neurological diseases (Amyotrophic lateral sclerosis (ALS) and Huntingtons Disease) as well as healthy controls. A range of statistical feature vector considered here from the Time-series gait data which are then reduced using correlation matrix. These feature vectors are then individually analysed to extract the best 7 feature vectors which are then classified using a Gaussian radial basis function kernel based Support vector machine (SVM) classifier. Results show that the 7 features selected for SVM achieves good overall accuracy of 83.33%, good detection rate for Parkinsons disease of 75% and low false positive results of 16.67%.