Predicting CH₄ Emissions from Rice Production in Southeast Asia via Feature Scaling and Polynomial Regression Modeling

This study investigates methane (CH<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">4</inf>) emissions prediction from rice production in Southeast Asian countries using feature scaling techniques and polynomial regression. The research focuses on rice production and emission data from ten Southeast Asian countries between 2003 and 2022. In addition, the data are obtained from the Food and Agriculture Organization (FAO) database. The feature scaling approaches namely min-max scaling and standard scaler are implemented and compared across different polynomial degrees to optimize prediction accuracy. It can be seen that a second degree polynomial regression model shows the best performance without feature scaling. However, the model performance improves after implementing feature scaling techniques with min-max scaling. The third degree polynomial regression with min-max scaling achieved optimal performance with R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> value of 0.8982 and MSE of 0.0099 demonstrating an effective balance between model complexity and prediction accuracy. Higher degrees of polynomials show marginal improvements while substantially increasing model complexity. This research provides valuable information for understanding and predicting agricultural methane emissions in Southeast Asia contributing to the development of more effective climate change mitigation strategies in rice producing regions.