Publication

Abstract : Individual movement analysis is becoming more open and common as sensors incorporated in mobile devices and machine learning algorithms progress. In the proposed framework, different classification techniques of machine learning algorithms such as Logistic Classifier and Support vector machine (SVM) are used to recognize human activities. The Fast Fourier transform (FFT) of tri-axial accelerometers that measure acceleration in all three spatial dimensions of smartphones is considered with different window sizes which correspond to the different number of samples from the sensor to predict the activities of an individual. The Performance analysis of individual classification for different window sizes is evaluated using accuracy, recall, precision, and F1-score. Results show that SVM performed better when compared to other model but there is a significant increase in accuracy as the window size increase for both the ML classification techniques.