AI-Powered Carbon Emission Prediction: A Deep Drive into LSTM Modeling

This study aims to compare the effectiveness of the Long Short-Term Memory (LSTM) model against the Random Forest Regressor for improving the accuracy of carbon emission monitoring. Two groups were tested in this study. Group 1 utilized an LSTM-based prediction model trained on real-world carbon emissions such as normalization, outlier removal, and feature engineering to enhance accuracy. Group 2 employed the Random Forest Regressor using the same dataset for comparison. The models were evaluated using accuracy, mean squared error (MSE), and statistical significance using a p-value threshold of 0.05 with a confidence interval of 95%. The results showed that the LSTM model achieved 94.00% while the Random Forest model achieved 85.00% accuracy. The statistical analysis confirmed that the improvement in accuracy was significant with a p-value < 0.05. The findings demonstrate that the LSTM model provides superior predictive accuracy and efficiency in real-time carbon emission monitoring, making it a reliable tool for sustainability efforts and environmental policy planning.