An Environmental Green Approach By Optimization Of Air Quality Index (AQI) Prediction Using Hybrid Machine Learning Combines With Swarm Intelligence Algorithm

Abstract

The pollution in the air is a one of the challenge of green environmental in urban areas that effects the natural human living. The local authorities interact to real-time monitoring and analyze the pollution data caused as per the present status of traffic situation in the city. As per this, it is essential to make appropriate precautionary measures and decisions accordingly. The Internet of Things(IoT) based sensors play a major role in dynamic of predicting air quality. In the city, due to increase of vehicle's traffic, industrialization, urbanization are cause the air contamination that effects the air pollution that effect breathing issue of human and cause health condition.  Air pollution is a major problem in the traffic areas of urban. Due of vehicles quantity and releasing various hazardous gases like CO, CO2, NO2, NO, SO2, NH3, PM2.5 cause environmental pollution that affects human health problems. Addition to this modern industrialization and heavy traffic contains harmful gas molecules are being spread and contaminated in the air. These particles in the size of PM10 and PM2.5 these are seriously effect the human health. In urban everyday releasing harmful gases and cause an un-curable diseases for human. In this paper, we performed pollution prediction by applying advanced regression methods actually predict air quality in the smart city environments. An optimization of Improved Grey Wolf Optimization(IGWO) combined with Decision Tree (DT) machine learning algorithm to find accurate values of prediction values of AQI  in the urban areas in India. The dataset is for experimental purpose is available with Root Mean Square Error (RMSE for real-time data of various city of India for the experimental purpose. The air pollution monitoring is an essential to predict to take necessary actions to manage the ecological balance in the city. In this paper apply regression method to predict the air pollution in the major cities. This proposed research experimental approach by verify the repository metrics via R-Square, Mean Absolute Error (MAE), Mean Squared Error(MSE), Root Mean Square Error (RMSE) metrics. The machine learning algorithms like K-nearest neighbor, Random forest regression, are comparing with the proposed algorithm. The proposed research Hybrid IGWO-DT algorithm providing  the best performance achievement results when compared with existing machine learning algorithms by  gaining maximize the accuracy levels of various cities in India.