International Journal of Science, Engineering and TechnologyAuthors: Hemlata, Jaishree Goyal, Manoj Pal
Abstract: Agriculture plays a crucial role in the economic development and food security of India. Predictable climate, soil variability, excessive fertilizer usage, and a lack of data-driven decision support systems often pose significant challenges for farmers in selecting suitable crops. The use of traditional farming practices is heavily dependent on experience and intuition, which may not always be in alignment with current environmental conditions and can result in low productivity and financial losses. This study presents a Smart Farming system that uses Machine Learning to accurately predict crops using key environmental and soil parameters such as Nitrogen (N), Phosphorus (P), Potassium (K), temperature, humidity, pH, and rainfall. A Random Forest classifier was trained using a dataset that is publicly available and contains 2200 samples and 22 crop classes. The proposed model's accuracy was approximately 96%, which demonstrates strong predictive performance for multi-class crop recommendations.
DOI: http://doi.org/10.5281/zenodo.21455972
