HYBRID MODEL AND FRAMEWORK FOR PREDICTING AIR POLLUTANTS IN SMART CITIES

Abstract

The pollution index of any urban area is indicated by its air quality. It also shows a fine balance is maintained between the needs of the populace and the industrial ecosystem. To mitigate such pollution in real-time, smart cities have a significant role to play. It's common knowledge that air pollution in a city severely affects the health of its dependents. More alarmingly, human health damage and disease burden are caused by phenomena like acid rain, and global warming. More precisely, lung ailments, CPOD, heart problems and skin cancer are caused by polluted air in congested urban places. Amongst the worst air pollutants, CO, C6H6, SO2, NO2, O3, RSPM/PM10, and PM2.5 cause maximum havoc. The climatic variables like atmospheric wind velocity, direction, relative humidity, and temperature control air contaminants in the air. Lately, numerous techniques have been applied by researchers and environmentalists to determine the Air Quality Index over a place. However, not a single technique has found acceptance from all quarters as being effective in every situation or scenario. Here, the main aspect relates to achieving authentic prediction in AQI levels by applying Machine Learning algorithms so worst situations can be averted by timely action. To enhance the performance of Machine Learning methods study adopted imputation and feature selection methods. When feature selection is applied, the experimental outcomes indicate a more accurate prediction over other techniques, showing promise for the application of the model in smart cities by syncing data from different monitoring stations.