Applications of Machine Learning Techniques on Air Pollution

In contemporary times, it is imperative to consciously monitor the nature of air quality in our environment on a continuous basis due to poor quality of air. This enables us to realize the commitments of global nations towards the SDG 3.9 goal to reduce air contamination by 2030 in a robust manner. This research paper aims to prognosticate with empirical models that are developed to envisage real time air quality index (AQI) for PM2.5, PM10, NO, NO2, NOx, NH3, CO, SO2, O3, benzene, toluene and xylene during pre-covid as well as post-covid era during the reference period 2015 to 2020. The seasonal auto-regressive integrated moving average time series model was applied to forecast the AQI with a model accuracy of test data greater than 90%. The research study investigates empirically the recurrent lapses over a long-term period compliance (2010-2020) regarding non-compliances with green norms of MOEFCC for air pollution levels from the NTPS (Narla Tata Rao thermal power station). The empirical results in terms of abatement of air effluence of thermal plants have proved that green technological interpellations are a laudable establishment of a regenerative economy.