Advanced Data Mining Techniques for Crop Yield Predictive Modeling: The Strategic Modified Gradient Boosting Regression (SMGBR) Approach

Agricultural productivity is critical for economic stability, particularly in agrarian regions like Banaskantha District, North Gujarat, where cabbage farming significantly contributes to local livelihoods. Effective crop yield prediction models are essential for planning, resource allocation, and sustainable agricultural practices. This study presents the Strategic Modified Gradient Boosting Regression (SMGBR) model, designed to improve the accuracy of cabbage crop yield forecasts. Unlike traditional Gradient Boosting Regression (GBR) models, the SMGBR model incorporates a dynamic learning rate adjustment, which enhances its ability to handle complex, non-linear relationships in agricultural datasets. The study utilizes an extensive dataset collected over twelve years (2010–2021), which includes critical variables such as historical crop yield records, detailed meteorological data, and soil health metrics. These factors are integrated into the model to refine predictive accuracy. The results indicate that the SMGBR model outperforms conventional approaches, delivering more reliable and precise yield forecasts. The improvements in prediction accuracy achieved by the SMGBR model are particularly beneficial for agricultural planners, policymakers, and farmers. This model can guide more informed decision-making on crop management, resource distribution, and long-term agricultural planning, ultimately promoting both economic and environmental sustainability. By offering precise, data-driven yield predictions, this research contributes to optimizing agricultural output while reducing waste and supporting the ecological balance in cabbage farming. The findings underscore the importance of advanced machine learning techniques in fostering resilience and sustainability in agricultural sectors.