Assessment of Human Activity Recognition based on Impact of Feature Extraction Prediction Accuracy
P. William, Govinda Rajulu Lanke, Dibyhash Bordoloi, Anurag Shrivastava, Arun Pratap Srivastavaa, Sheetal Vishal Deshmukh
Abstract
Recognition of human activities by analyzing smartphone data which is being collected via accelerometer and gyroscopic sensors has been a critical area of research and it has been providing solutions to various real-world problems in various domains like healthcare and others. For accurate prediction of human activities, the data is collected using accelerometer and gyroscopic sensor from a smartphone and a feature vector of size 561 is created. A set of features is calculated over this data. In this paper, a systematic analysis of these features is being done and an extensive result on how the choice of features affects the recognition accuracy for various human activities is being provided.