Machine Learning-Based Fertilizer Recommendation System for Sustainable Crop Production

The efficient use of fertilizers plays a critical role in enhancing crop yield while minimizing environmental degradation. Traditional fertilizer practices often result in the overuse or underuse of key nutrients such as nitrogen (N), phosphorus (P), and potassium (K), leading to reduced productivity and soil health deterioration. This study presents a machine learning-based fertilizer recommendation system designed to predict optimal NPK values based on crop type and soil parameters. The model is trained on a curated dataset compiled from agronomic guidelines issued by the Food and Agriculture Organization (FAO) and the Indian Council of Agricultural Research (ICAR) [1][2]. Key features include soil pH, moisture, and crop species, with the Random Forest algorithm demonstrating the highest accuracy in nutrient prediction. The model consistently achieves high performance in classifying nutrient needs across diverse crops, ensuring tailored recommendations that align with sustainable farming practices. This approach empowers farmers with a data-driven tool to apply fertilizers precisely, reducing input costs and enhancing crop health. The system was implemented and tested using Python and Google Colab, providing a scalable and accessible platform for practical deployment in agricultural decision-making. Keywords: Machine learning, fertilizer recommendation, NPK prediction, soil nutrient analysis, crop-specific nutrient management, sustainable agriculture, Google Colab implementation, precision farming, agronomy, data-driven decision making.