Identification of Human Movement Through a Novel Machine Learning Approach

Abstract

This study aims to identify human movement activities using machine learning, focusing on a novel ensemble approach. The objectives are twofold: to apply a new machine learning method for activity recognition and to outperform recent approaches in accuracy. A standard dataset from Kaggle contained six activities: standing, sitting, lying, walking, walking downstairs, and walking upstairs. The study used an SVM and Logistic Regression ensemble alongside other standard classifiers. The proposed ensemble achieved an accuracy of 95.45% on the test data, surpassing other models. This research shows the potential of the ensemble approach for accurate human movement identification with potential uses in various domains.