Data-Driven Crop Yield Prediction Using Machine Learning for Sustainable Agricultural Development

The Indian economy is predominantly agrarian, with Karnataka playing a significant role in national food production. Accurate crop yield prediction is essential for data-driven decision-making, efficient resource utilization, and the advancement of sustainable agricultural practices. Traditional statistical methods and standalone machine learning models often struggle to capture the complex, nonlinear, and time-varying interactions among crop characteristics, soil properties, and climatic conditions. To address these challenges, this paper proposes SmartAgri-Net, a hybrid ensemble architecture that integrates Long Short-Term Memory (LSTM) networks for advanced temporal modeling, Gated Recurrent Units (GRU) for efficient sequential learning, and Decision Trees (DT) for rule-based interpretability. The model is trained using the Raitamitra dataset provided by the Government of Karnataka, covering four major crops—rice, maize, ragi, and groundnut. The dataset includes attributes such as crop type, cultivation area, production volume, weather conditions, geographical location, and soil nutrient content. The proposed framework incorporates systematic data preprocessing, outlier detection, normalization, exploratory data analysis (EDA), feature selection, and ensemble model construction. SmartAgri-Net employs a weighted ensemble strategy to combine the predictions of DT, GRU, and LSTM models, thereby improving predictive accuracy and generalization capability. Experimental results demonstrate that the proposed model outperforms individual baseline models across multiple evaluation metrics, including Mean Absolute Error (MAE), Root Mean Square Error (RMSE), Coefficient of Determination (R²), and Mean Absolute Percentage Error (MAPE). The framework supports precision agriculture and informed policymaking by delivering a scalable, accurate, and interpretable crop yield prediction system. Future work will focus on integrating real-time IoT-based agricultural data and developing region-specific crop recommendation systems to further enhance sustainability and resilience.