Parkinsonian Gait Motor Impairment Detection Using Decision Tree

In this study the potential of decision tree (DT)specifically Classification and Regression Tree (CART) isinvestigated in classifying Parkinsononian gait pattern due to motor impairment. Firstly, gait features extracted duringwalking experiments namely basic spatiotemporal parameters,kinetic parameters and kinematic parameters by twelve PDpatients and twenty control group are acquired as database.Next, CART is chosen as classifier for identifying the distinct features between these two groups. Initial results attained proven that three kinematic gait features namely maximum flexion of ankle angle, maximum extension of knee angle as well as maximum extension of hip angle contributed as distinct features for detection of motor impairment in PD patients due to perfect success rate attained. Thus, these three distinct features could be used for early diagnosis of Parkinsonian gait.