Abstract
In previous studies, we proposed a system for classifying gait based on step length and foot-ground clearance using arm acceleration. In the present study, we evaluated the application of machine learning to gait classification. The method was tested empirically on the classification of three gait patterns performed by 10 young and healthy participants. The three gait patterns were normal step, high step, and long step. Using measures of accuracy, precision, recall, and F-measure, we compared the performances of the following six classifiers: naive Bayes, support vector machine, neural network, logistic regression, instance-based classifier, and decision tree. The proposed method was shown to be capable of classifying the three gait patterns of seven participants with an accuracy greater than 0.6. This suggested that the proposed machine learning-based method is appropriate for its application in gait classification systems.
