Implementation of Smartphone Based Activity Recognition

Hobby reputation is a detectable technique of recognizing distinct sports finished by way of an man or woman in exclusive conditions. Such sports consist of walking, jogging, strolling, going up stairs, taking place stairs, mendacity down. Our main objective is to access various acts taken by individual. Although there are a variety of approaches for detecting various activities, we prefer to employ Smartphone accelerometers due to their widespread availability and ease of use. There are a diffusion of open-supply sensor-based completely time-series datasets that are used to check rule overall performance, that is finally applied to locate human pastime. We have balanced dataset with information and labels are essential for testing the performance of a set of rules. Size, labelling, accuracy, a wide range of activities carried out by each user, formed sensor information contained, and a variety of other factors all influence the size, labelling, and accuracy of such datasets. In this study, we evaluate the popular datasets. We have performed different algorithm i.e. Decision Tree, Logistic Regression, SVC, Gradient Boosted Decision Tree, Random Forest alongside this we have also predicted precision, recall, and F1 score. On the basis of the above parameter the accuracy rate achieved by Logistic Regression in UCI-HAR is 96.19% and in WISDM is 100%