Fertilizer Management with Machine Learning: A Farmer’s Guide to Improved Yield

This paper addresses the critical issue farmers face in selecting the appropriate type and amount of fertilizer to optimize crop productivity and profitability. Leveraging machine learning techniques, specifically a K-Nearest Neighbors (KNN) regression model, the study focuses on recommending nutrient (N, P, K) quantities per parcel to enhance crop yields. Evaluation of the model demonstrates promising results, with mean absolute error (MAE) values of 0.080 for N, 0.059 for P, and 0.025 for K, and corresponding mean squared error (MSE) values of 0.115, 0.080, and 0.066, respectively. These findings offer farmers reliable guidance in planning their agricultural activities, facilitating informed decisions regarding fertilizer quality and quantity.