Novel approach for spastic hemiplegia classification through the use of support vector machines.

The use of computer based classification methods for spastic hemiplegia (SH) is a relatively new approach. Before and still, the classification is primarily done by specialized physicians through manual template matching on kinematics plots of the sagittal plane data from the affected limb generated by gait laboratories. In the past, several methods for computer based automation of this classification have been attempted. This paper introduces the use of support vector machine (SVM) as a model that contributes to this process. The results obtained from the use of SVM are quite efficient considering the data set utilized involves patients from an array of ages and both sex. It should be mentioned that even though the number of patients is not trivial; the percentages of accurate classification represented are promising. Since the SVM method improves its accuracy with an increased number of training cases, this approach has the advantage of becoming more accurate with time. Many more significant discoveries are expect from the introduction of SVM in the analysis of SH and further studies of gait related pathologies are expected to contribute as well.