Abnormal GAIT classification using hybrid ELM

With a mounting increase in gait diseases, a numerous researches are being carried out on gait analysis to determine any abnormality in gait. Machine learning techniques can be used here as it learns like intelligence and covers a wide range of processes that are complicated to classify accurately. Among the existing machine learning techniques, Extreme Learning Machine(ELM) algorithm serves better and can classify the data precisely. In this paper, a two phase algorithm is proposed and implemented to detect the abnormal gait. In the first phase, ranking is performed to determine the top gait features. This paper uses T-Test technique for this purpose. In the second phase, Machine learning algorithms are used for training and testing the occurrence of abnormal gait. For this purpose, this paper uses a modified version of Extreme Learning Machine called Hybrid Extreme Learning Machine (HELM). HELM uses the Analytical Network Process (ANP) for choosing the input weights and hidden biases. The proposed technique is evaluated using CGA Normative Gait database. Experimental results prove that the proposed technique for gait classification results in better accuracy compared to the existing techniques.