Leveraging nonlinear deep learning models for intelligent crop recommendation in precision agriculture

Abstract For countries with developing economies, agriculture, particularly the farming and livestock business, plays an essential role in their economic growth. Agriculture is the basis of food security and supplies the farming labor and means of subsistence for the population of the city and rural areas. Modern technologies in agriculture ensure timely, informed, and precise decisions concerning agriculture and related sectors along with the seasonal and climatic variations. This paper suggests a method for selecting the right crop that makes use of a number of deep learning-based crop recommendation systems and allows crop selection for a particular area based on soil and climate. The method incorporates various deep learning techniques, including ANNs, CNNs, RNNs, LSTMs, and CRNNs, in crop categorisation. Individual classifier findings are then integrated with an ensemble classifier to create more accurate crop recommendations by leveraging the complicated nonlinear interactions between diverse agro-environmental parameters. The experimental results reveal that the proposed ensemble model outperforms individual deep learning models, with an overall accuracy of 96.95%. The findings indicate that deep learning-based crop recommendation systems can assist farmers optimise crop selection, support crop selection and improve data-driven agricultural decision-making under varying soil and climatic conditions.