Classification of Gait Parameters in Stroke with Peripheral Neuropathy (PN) by using k Nearest Neighbors (kNN) Algorithm

This paper presents the gait pattern classification between 3 groups which are control, stroke only and stroke with Peripheral Neuropathy (SPN) using k-Nearest Neighbors (kNN) algorithm. Control group has been used as a reference or baseline in order to see the difference in the gait pattern. The model able to classify patients into their respective group based on the gait parameters collected. Furthermore, the findings also will help them to monitor patient’s performances in rehabilitation program from time to time. 29 subjects has been recruited (9 SPN, 10 stroke subjects and 10 control subjects) with range of age between 40 to 65 years old. Additionally, all subjects must be able to walk freely without any cane or mechanical aid during walking. Vicon® Nexus Plug-in-Gait has been used to compute the kinematic gait parameters. From the results, it is found that there are 9 significant differences in kinematic angles and spatio-temporal data. The classification model developed has been successfully discriminate three different groups with 83.33% accuracy.