Document Type : Research Paper

Authors

• D Saravanan ¹
• K Santhosh Kumar ²

¹ Department of CSE, IFET College of Engineering, Villupuram, India.

² Department of IT, Annamalai University, Chidambaram, India.

Abstract

In recent decades air pollution and its associated health risks are in growing numbers. Detecting ‎air pollution in the environment and alarming the people may accomplish various advantages ‎among health monitoring, telemedicine, and industrial sectors. A novel method of detecting air ‎pollution using supervised learning models and an alert system using IoT is proposed. The main ‎aim of the research is manifold: a) Air pollution data is preprocessed using the feature scaling ‎method, b) The feature selection and feature extraction process done followed by performing a ‎Recurrent Neural Network and c) The predicted data is stored in the cloud server, and it provides ‎the end-users with an alert when the threshold pollution index exceeds. The proposed RNN ‎reports enhanced performance when tested against traditional machine learning models such as ‎Convolutional Neural Networks (CNN), Deep Neural Networks(DNN), and Artificial Neural ‎Networks(ANN) for parameters such as accuracy, specificity, and sensitivity.‎

Keywords

• Internet of Things
• Convolutional Neural Networks (CNN)
• Deep Neural Networks ‎‎(DNN)
• and Artificial Neural Networks (ANN)
• Recurrent Neural Network