Optimized Crop Yield Forecasting Using the Naive Bayes Regression Algorithm in Smart Agriculture

Crop yield forecasting is significant for ensuring food security, optimizing resource utilization, and aiding decision-making in smart agriculture. Traditional methods such as Multiple Linear Regression (MLR), Support Vector Regression (SVR), and Random Forest Regression (RFR) have gained popularity for yield forecasting. However, such methods commonly struggle with noisy datasets, feature correlation, and intricate non-linear relationships, resulting in compromised predictive accuracy. To address such shortcomings, this study introduces the Optimized Naïve Bayes Regression Algorithm (ONBRA) to predict crop yield. ONBRA incorporates the probabilistic advantages of Naïve Bayes and optimizes the selection of advanced features and smoothing. This algorithm offers enhanced predictive accuracy under varying agricultural conditions. The experimental results show that ONBRA improves forecasting accuracy by 6.8%, the Mean Absolute Error (MAE) by 5.3%, and the R-squared (R²) by 7.1% compared to traditional methods. The results confirm that the optimized algorithm handles climatic, soil, and crop variability better than traditional methods, providing a better and sustainable solution for smart farming.