Human activity recognition based on AdaBoost ensemble classifier

Abstract

Human activity recognition (HAR) has been widely applied in the field and has good application prospects. Various classifiers in machine learning have shown excellent performance in their own fields. In this paper, AdaBoost ensemble classifier for human activity recognition is proposed to improve the performance of human activity recognition technology by using a weighted combination of multiple classifiers. The experimental results of HAR data were evaluated, and the total classification accuracy and receiver operating characteristic (ROC) area were calculated. The results show that the AdaBoost ensemble classifier framework proposed in this paper can accurately identify six kinds of human activities, and the AdaBoost ensemble classifier algorithm can significantly improve the HAR recognition accuracy.