Study and Analysis of Supervised Machine Learning Techniques for Human Activity Recognition and Its Implementation Using Smartphones Sensors

In this paper, we modeled a system to recognize human activities using integrated sensors like gyroscopes and accelerometers in smartphones. To perform human activity recognition (HAR) accurately, appropriate machine learning algorithms and dimensionality reduction methods are utilized. The feature reduction techniques like principal component analysis (PCA) and correlation analysis (CA) are also used. This paper mainly concentrates on the recognition of user’s actions by utilizing different machine learning classification approaches like k-nearest neighbor, logistic regression, artificial neural network, decision tree, and Naïve Bayes algorithms. The best result of the research is ANN with a classification accuracy of 97% via the use of the CA feature reduction technique.