APD - ML: Air Pollution Detection Using Machine Learning Algorithms

Abstract

To analyze the air quality of any country, a machine learning technique is being developed and an air quality indicator is proposed for a particular area. Air Quality Index is considered to be a basic measure which can indicate the levels of SO2, NO2. etc. over a particular amount of time. We technologically put forward a model to determine the air quality index in view of historical data of preceding years and computing the same for the forthcoming year considering it as a gradient decent attached boosted multivariable regression problem. We enhance the proposed model's effectiveness by relating cost estimation on behalf of the problem to be a predictive one. Thus this proposed system resolve successfully and work well to envisage the air quality indicator of any entire country or state or any bounded region furnished with enough historical data about contaminants in air. In the proposed model, subsequently machine learning technique is assimilated, upright enactment with performance is accomplished further than the standard regression model. The implementation of envisaging air quality index is prepared for our country India as well as accurateness of 96% is attained via XG Boost Algorithm joined with LightBGM algorithm to find an accurate solution that is in adjacent proximity to the ideal solution.