Recently, with the development of machine learning theory,
a wide variety of novel learning and classification algorithms
have been more and more prevailing in analyzing quantitative
gait data. Especially, some advanced learning algorithm for the
automated recognition of gait change has attracted much
attention because they can solve classification problems with
better performance. For example, Support Vector Machine
(SVM), as a powerful classification technique, has been
employed to classify young and elderly gait pattern for the
assessment of the change of gait function[2,3, 4]. The present
studies have demonstrated that it is important to improve gait
classification performance for evaluating the gait change
exactly. In gait classification algorithms, as we know, the gait
patterns to be classified are usually required to represent as
points in a high-dimensional feature space. In order to improve
the gait classification performance, it is required to extract
some significant gait features from the initial gait features
space for reducing the redundant information before the gait
classification algorithm is executed. In fact, the interaction
between gait variables is a complex non-linear fashion because
of the intrinsic non-linear dynamics of human movement. So,
the key step in the gait classification algorithm designed for
Recently, with the development of machine learning theory,a wide variety of novel learning and classification algorithmshave been more and more prevailing in analyzing quantitativegait data. Especially, some advanced learning algorithm for theautomated recognition of gait change has attracted muchattention because they can solve classification problems withbetter performance. For example, Support Vector Machine(SVM), as a powerful classification technique, has beenemployed to classify young and elderly gait pattern for theassessment of the change of gait function[2,3, 4]. The presentstudies have demonstrated that it is important to improve gaitclassification performance for evaluating the gait changeexactly. In gait classification algorithms, as we know, the gaitpatterns to be classified are usually required to represent aspoints in a high-dimensional feature space. In order to improvethe gait classification performance, it is required to extractsome significant gait features from the initial gait featuresspace for reducing the redundant information before the gaitclassification algorithm is executed. In fact, the interactionbetween gait variables is a complex non-linear fashion becauseof the intrinsic non-linear dynamics of human movement. So,the key step in the gait classification algorithm designed for
การแปล กรุณารอสักครู่..
