A Machine Learning Model for Predicting Carbon Emission

Abstract

Air pollution impacts human health in various ways, including by depleting the ozone layer. This study aimed to utilise available data to develop a machine-learning model that predicts carbon emissions. The dataset was processed, converted to a time series, and split into training and test sets at a 70:30 ratio. The Long Short-Term Memory (LSTM) and Autoregressive Integrated Moving Average (ARIMA) models were employed to develop the model. Root Mean Squared Error (RMSE) metrics were used to evaluate the results. The findings indicate that applying the LSTM model to a large dataset with a high number of epochs yields better accuracy than using ARIMA on the same dataset. The LSTM achieved a lower RMSE of 0.0440 and better predicted carbon emissions than ARIMA. The system developed is recommended for countries, organisations, and agencies to monitor carbon-related air pollution.