Loading... Loading...
Grenze Logo
GRENZE International Journal of Engineering and Technology Vol. 12 (2026), Issue 1

AgroAdviser: A Multi-Model Framework Integrating DeepFM with Random Forest and Gradient Boosting for Smart Crop Yield and Fertilizer Prediction

Authors

Reshma Begum, Mrutyunjaya. S. Yalawar

Abstract

AgroAdvisor integrates crop yield prediction, crop recommendation, and fertilizer recommendation using a hybrid Random Forest with Extreme Gradient Boosting (RFXGB) and Deep Factorization Machine (DeepFM) approach. By analyzing historical data, soil conditions, and weather patterns, it enhances decision-making with high accuracy. RFXGB ensures efficient feature selection, while DeepFM captures complex feature interactions, outperforming traditional models like SVM, XGBoost, and ANN. The system provides actionable insights for optimal crop selection, yield forecasting, and fertilizer application, improving productivity and sustainability. Its modular design allows seamless integration of features like weather forecasting and pest control, ensuring scalability and adaptability for modern agriculture.