Multilayer machine learning algorithm to classify diabetic type on knee dataset

since last decade, the diabetes risks are increasing in children and adults. Various approaches have been proposed for early detection of the diabetes and prevention on it. Some methods use EMG signals for diabetes classification, due to motion artifacts in the EMG signals during acquisition of signal, these approaches are not able to classify the signal efficiently. To overcome this we propose anew method by considering time domain and frequency domain features of the EMG signals and to perform the classification we use neural network. This method is executed using MATLAB tool and simulation study shows the accuracy of proposed approach is 97.05%.