Leveraging Ensemble Learning Technique for Efficient Fertilizer Recommendation

Gimba, Ahmed Mohammed
Pradeep Kumar Mishra

JITSI : Jurnal Ilmiah Teknologi Sistem Informasi
Vol 6 No 4 (2025)

Agricultural productivity plays a critical role in improving crop yields, and the suitable use of fertilizers plays a significant role in enhancing crop yield. Traditional fertilizer recommendation approaches often rely on generalized strategies that may not account for discrepancies in soil properties, climatic conditions. To address this limitation, we proposed an intelligent Fertilizer Recommendation System (FRS) using an Ensemble Learning method. This system integrates multiple ensemble learning models, such as Bagging, AdaBoost, GBoosting, Extra Trees, and CatBoost to enhance recommendation accuracy. The ensemble model is trained on soil parameters (N) nitrogen (P) phosphorus, (K) potassium, and moisture to recommend the optimal fertilizer type in Andhra Pradesh region, India. The result shows that all ensemble models utilized were effective, and CatBoost model has achieved 94.78% with highest accuracy, when compared with the other ensemble models.