OEL: Optimized Ensemble Learning for Datadriven Rice Yield Prediction

xRice is one of the staple food item for a substantial people in the universe, calculating perfect predictions of its production crucial for food security and agricultural planning. This study leverages Machine Learning (ML) approaches to predict rice production using a samples of historical agricultural data. Five ML models Linear Regression (LiR), Polynomial Regression (PoR) with Lasso Regularization, Random Forest (RFR), Gradient Boosting (GBR), and Support Vector Regression (SVR) were implemented with optimal parameters using GridSearch Cross-Validation (GridSearchCV). This research work proposed correlation coefficient feature selection and ensemble technique to predict the rice yield. The rice yield case study has been conducted using the cultivation data of a few districts in Andhra Pradesh. The models have been assessed based on key measures such as <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{R}^{\mathbf{2}}$</tex> Score, Mean Squared Error (MSqE), and Mean Absolute Error (MAbE) to prove the effectiveness. The ensemble approach resulted an optimal performance with an <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{R}^{\mathbf{2}}$</tex> score of 0.9891 compared to individual ML models. The findings underscore the importance of combining data-driven ensemble learning in agricultural farming to improve yield predictions, optimize resource allocation, and support sustainable farming.