CO2 emissions prediction from coal used in power plants using univariate and multivariate machine learning models

Abstract

Abstract Carbon dioxide (CO2 ) emission has increased rapidly due to the predominant usage of fossil fuels. The energy sector contributes a considerable amount towards the total share that belongs to CO2 emissions worldwide. In this work, we have applied the Multivariate and Univariate variants of time-series, machine learning and deep learning models over the CO2 emissions dataset. The dataset is collected central electricity authority containing the attributes as coal supply information, CO2 emissions, peak demand, and peak met. The performance of the applied models is tested using performance metrics such as RMSPE, MAE, RMSE, MSE, MAPE, SMAPE, and RAE. The dataset is collected from 2005-2021 to perform the test and train, and further, we have forecasted the CO2 emission from 2022-2050 by applying the best performing models. The findings of the work show that autoregression is the best-performing model and obtained the best rank i.e. 1.85 by applying the Friedman ranking. A comparative study is also done over multivariate and univariate analysis.