A Dream City: Identifying Red Spots Based On IoT Based Air Pollution Prediction Model

Abstract

The rapid growth of factories severely increase the air pollution with various particulates. Even though every country insists standards for the emission of pollutants, the violation happens continuously. The identification of violation factories is very essential to save the earth. In this paper, we presented a novel Air pollution free Dream City (APFDC) framework which is based on Nonlinear autoregressive neural network along with Levenberg-Marquardt neural optimizing algorithm for the prediction of factories who violate the standards of pollution control board. We process and analyze obtained IOT based BIG data by means of neural network and predicted accurate violated factories as possible. Obtained results from prediction are then optimized by iteration method designed for finding the best possible combination of neural network parameters. Our proposed model pulls out the air pollution severity and provides the guideline for the requirement of strict supervising. We used city pulse database which consist of 8 features including ozone, particulate matter, carbon monoxide, sulfur dioxide, nitrogen dioxide, longitude, latitude and timestamp for the right prediction. The acquired experimental result showed that the proposed method performs better than conventional methods.