Performance Analysis of Machine Learning Models for Air Pollution Prediction

Abstract

Air pollution includes contamination of air due to harmful gases, residues, fumes, etc. Contaminated air gives rise to important issues for the solid endurance of plants, organisms and individuals, including natural life. This paper focuses on predicting air pollutants using machine learning (ML) techniques and its performance analysis. Various regression and classification models like Support Vector Machine (SVM), Random Forest Classifier, Logistic Regression, Linear Regression and Random Forest Regression are used to optimize the air pollutants for better accuracy in forecasting. The performance of ML models is evaluated using State Pollution Control Board (SPCB) dataset, Odisha. The performance of Regression models is evaluated using Root Mean Square Error (RMSE) and Mean Absolute Error (MAE). It prevails in Random Forest Regression having RMSE and MAE as 2.63 and 3.32 respectively. For classification models, Random Forest Classifier precede with an accuracy of 93.5 %. The efficient performance of the model in predicting air pollutants can help in alerting the public to safer living.