Sensor Data Analysis and Management

6 Deep Learning Analysis of Location Sensor Data for Human-Activity Recognition

Hariprasath Manoharan¹, Ganesan Sivarajan², and, Subramanian Srikrishna³

¹ Assistant Professor, Department of Electronics and Communication Engineering, Audisankara College of Engineering and Technology, Gudur, Andhra Pradesh² Associate Professor, Department of Electrical and Electronics Engineering, Government College of Engineering, Salem, Tamil Nadu³ Professor, Department of Electrical and Electronics Engineering, Annamalai University, Chidambaram, Tamil Nadu

6.1 Introduction

For improving the working efficiency of sensors and for testing them under different conditions, a predictive algorithm is essential. This is possible only when deep learning methods are used, where different strategies are followed when any problem occurs on the network. If sensors are installed, then the network depends on main node for the purpose of storing and accessing the data. For sensing the information and sending it to applications, such as those for health monitoring, there should be less delay. This robust prediction of health is necessary because it can save people’s lives. There are many deep learning models integrated with neural networks, where many hidden layers are included in the shape of a cascade.

This cascaded network provides the necessary machine learning techniques for the automatic detection of datasets. Most of the useful properties from datasets are based on soft target models with probability ...