Performance Comparison of Activity Recognition Classifiers using Big Dataset

Abstract

Human activity recognition is a promising concept of pervasive computing. Multiple number of on body sensors is employed to achieve this task. Activity Recognition Chain (ARC) makes the process of activity recognition possible. ARC includes various stages namely, data acquisition, preprocessing, segmentation, feature extraction, classification, and decision fusion. Amongst these, classification is the most critical stage. The paper deals with classifying human activities on a big dataset. The classifiers include Naive Bayes, HMM, DA, and k-NN. The paper shows which classifier is best suited in big data environment for classifying the activities.