AI BASED FERTILIZER RECOMMENDATION SYSTEM FOR CROP OPTIMIZATION

This paper presented an AI-Based Fertilizer Recommendation System for Crop Optimization that addresses a critical gap in accessible agricultural decision-support tools. The system's rule-based approach provides transparent, interpretable recommendations grounded in agronomic science, making it particularly suitable for non-technical users such as smallholder farmers and field agents. The client-server architecture separates the user interface from backend logic, enabling independent development, maintenance, and scaling of each component. Docker containerization ensures the system can be deployed on any platform — from local machines to cloud infrastructure — without dependency conflicts. Our tests showed that the system is highly effective at providing accurate fertilizer recommendations across a wide range of soil conditions. Measurements including recommendation accuracy, API response time, and interface usability all confirm that the system is dependable and performs as intended. The use of a rule-based decision engine ensures that recommendations are explainable, which increases farmer trust and adoption. However, the study also identified key limitations: reliance on manual soil parameter input, absence of environmental  data  integration,  internet  connectivity requirements, and lack of multi-language support. These areas represent clear directions for future improvement. Future enhancements could include IoT sensor integration for automated real-time soil data collection, a lightweight offline mobile application, incorporation of weather and seasonal data into recommendations, multi-language localization, and a machine learning layer that learns from historical recommendation outcomes to continuously improve accuracy. Since inefficient fertilizer use is a growing concern for food security and environmental sustainability, effective recommendation systems are critically needed. This project lays the groundwork for more advanced research to make agricultural AI tools more accurate, accessible, and impactful in real-world farming contexts.

[1]  C. Musanase, E. Niyigaba, and F. Nkurikiyeyezu, "Data-Driven Analysis and Machine Learning-Based Crop and Fertilizer Recommendation System for Revolutionizing Farming Practices," Agriculture, vol. 13, no. 5, 2023. [2]  R. Akhter and S. A. Sofi, "Precision Agriculture Using IoT Data Analytics and Machine Learning," J. King Saud Univ. – Comput. Inf. Sci., vol. 34, no. 8, pp. 5602–5618, 2022. [3]    G. Rajamanickam and G. Mani, "Soil Fertility Prediction Using Machine Learning Models," Applied Soft Computing, 2020. [4]  R. Katarya et al., "The Impact of Machine Learning Approaches in Diverse Stream," in Proc. 3rd Int. Conf. Computing Methodologies and Comm. (ICCMC), IEEE, 2019. [5]  Q. Chen, H. Zhang, and L. Wang, "Cloud-Based IoT System for Real-Time Crop Health Monitoring and Fertilizer Application," Sensors, vol. 23, 2023. [6]  Government of India, "Soil Health Card Scheme," Ministry of Agriculture and Farmers Welfare, 2024. [Online]. Available: https://soilhealth.dac.gov.in [7]    Kaggle, "Crop Recommendation Dataset," 2023. [Online]. Available: https://www.kaggle.com/atharvai ngle/crop-recommendation-dataset [8]    A. Kamilaris and F. X. Prenafeta-Boldu, "Deep Learning in Agriculture: A Survey," Comput. Electron. Agric., vol. 147, pp. 70–90, 2018. [9]  F. Pedregosa et al., "Scikit-learn: Machine Learning in Python," JMLR, vol. 12, pp. 2825–2830, 2011. [10]   S. Chakraborty and S. Bhunia, Smart Agriculture Using Deep Learning and IoT. Singapore: Springer Nature, 2022.