in order to assess the change of human
gait function. For instance, the artificial neural network (ANN)
based on machine learning algorithm has been employed for
the automated recognition of gait pattern change using the
different kinds of gait data[1,2]. Holzreiter and Kohle applied
ANN to identify normal and pathological gait by using kinetics
data[1]. Lees and Barton used ANN to discriminate the
different gait patterns based on gait features extracted from hip
and knee joint-angle measures. As we know, in gait
classification algorithms, the classification performance mainly
rely on the extracted or selected gait features from the initial
gait variables, that is, pre-processing for initial gait variables is
the first important step for improving classification
performance[3,4].