Machine learning-based Time Series Models for Effective CO2 Emission prediction in India

Abstract As per the report of datacommons.org, CO2 emission in India is 1.80 metric tons per capita, which is very harmful for living beings. So, this paper presents India's detrimental CO2 emission effect with the prediction of CO2 emission for the next ten years based on univariate time series data. We used three statistical models i.e., AutoRegressive Integrated Moving Average (ARIMA) model, Seasonal AutoRegressive Integrated Moving Average with eXogenous factors (SARIMAX) model, and Holt-Winter model, two machine learning models, i.e., Linear Regression and Random Forest model and one deep learning model, i.e. Long Short Term Memory (LSTM) model. The performance analysis shows that LSTM, SARIMAX, and Holt-Winter methods are accurate models and have the Least Mean Squared Error(MSE), Root Mean Square Error (RMSE), and Median Absolute Error (MedAE) for this kind of univariate data distribution.