Stacked Ensemble Machine Learning Techniques based Predictive Modelling of Crop Yields

Agriculture is crucial for food security and economic stability in India, where a most of the population relies on farming. Accurate crop yield prediction is essential for informed planning, efficient resource allocation, and maximizing agricultural productivity. This paper proposes a novel approach for predicting crop yield using an ensemble learning model. The proposed system utilizes historical agricultural data—including district, crop_year, season, area, and production for Tamil Nadu. The stacking ensemble model proposed in this paper integrates K-Nearest Neighbors Regressor and Multiple Linear Regressor as base learner and Decision Tree Regressor as the meta-learner. This ensemble approach enhances prediction performance by leveraging the strengths of each individual model while minimizing their weaknesses. Experimental results, evaluated using R-squared (R 2 ) metrics, show that the Stacked Ensemble Regressor outperforms standalone models in terms of accuracy. This system offers strong decision-making support for farmers and agricultural stakeholders, helping them make informed, data-driven choices that enhance sustainability and efficiency in farming.