This research presents an advanced Explainable AI-based crop yield prediction and advisory system developed for precision agriculture. The system utilizes the XGBoost algorithm to achieve high prediction accuracy and integrates SHAP-based interpretability techniques to provide transparent and understandable insights into model decisions. The model analyzes key agricultural parameters such as soil nutrients (Nitrogen, Phosphorus, Potassium), temperature, humidity, pH, and rainfall to predict crop yield effectively. In addition to prediction, the system generates actionable recommendations for farmers, helping them optimize crop productivity and manage soil health efficiently. The developed system includes an interactive dashboard interface and automated PDF report generation, making it practical for real-world agricultural applications. This approach bridges the gap between artificial intelligence and farmer usability by transforming complex predictions into clear and actionable insights.
Deepthi Chandra Gandu (Tue,) studied this question.