Random Forest–Based Integrated Crop Recommendation and NPK-Driven Fertilizer Decision Support System for Precision Agriculture

Precision agriculture calls for data-driven methods to determine the optimal choice of crops and fertilizers depending on soil conditions and climate. In this paper, an integrated decision support system is presented, which is based on Random Forest crop classification and NPK-based fertilizer recommendation system. The proposed system works with seven agronomic features, including nitrogen, phosphorus, potassium, temperature, humidity, soil pH, and rainfall, from crop recommendation dataset with 2,200 entries and 22 classes of crops. After processing data and separating them into training set (80%) and testing set (20%), a Random Forest classifier with 100 decision trees was built to classify the crops. The classified crop is matched with the nutrient knowledge base, and the current soil NPK content is compared to the nutrient needs of the crop. Finally, the rules-based decision engine finds out nutrient deficiencies and makes fertilizer recommendations. The performance of the classifier on the test dataset was 97.6% accuracy, 97.1% precision, 96.8% recall, and 96.9% F1-score.