SVM Models in the Diagnosis of Balance Impairments

Abstract

Trip-related falls is a major problem in the elderly population and research in the area has received much attention recently. The focus has been on devising ways of identifying individuals at risk of sustaining such falls. The main aim of this work is to explore the effectiveness of models based on support vector machines (SVMs) for the automated recognition of gait patterns that exhibit falling behaviour. Minimum foot clearance (MFC) during continuous walking on a treadmill was recorded on 10 healthy elderly and 10 elderly with balance problems and with a history of tripping fails. MFC histogram characteristic features were used as inputs to the SVM model to develop relationships between MFC distribution characteristics and healthy/balance-impaired category. The leave-one-out technique was first utilized for training the SVM model in order to discover the appropriate choice of kernel. Tests were conducted with various kernels (linear, Gaussian and polynomial) and with a change in the regularization parameter, C, in an effort to identify the optimum model for this gait data. Then using a two-fold cross-validation technique, the receiver operating characteristics (ROC) plots of sensitivity and specificity were further used to evaluate the diagnostic performance of the model. The maximum accuracy was found to be 95% using a Gaussian kernel and the maximum ROC area = 0.88, when the SVM models were used to diagnose gait patterns of healthy and balance-impaired individuals. These results suggest considerable potential for SVM-based gait classifier models in the detection of gait changes in older adults due to balance impairments and falling behavior. These preliminary results are also encouraging and could be useful not only in the diagnostic applications but also for evaluating improvements or otherwise in gait function in the clinical/rehabilitation contexts