Intelligent Fertilizer Management System Using Optimized Artificial Neural Network Approach for Enhance Prediction

Abstract

In agriculture, maximizing crop yields and advancing sustainable farming methods depend on efficient fertilizer management. Artificial intelligence (AI) devices have created tools in agriculture to assist farmers in obtaining precise and regulated cultivation. The right way to predict fertilizer is more important in order to satisfy farmer demands, improve yield output, and manage agricultural operations. This research uses deep learning (DL) models to provide a new introduces an innovative Automated Fertilizer Management System (AFMS) employing an Optimized Artificial Neural Network (OANN) model, enhanced through back propagation and chain rule techniques. The suggested approach makes use of sophisticated optimization techniques, including hyper parameter tuning and regularization, to refine the back propagation process and improve learning efficiency. By leveraging the chain rule for gradient computation, the model ensures accurate and efficient weight updates during training. The ANN is trained on a diverse dataset that includes soil nutrient levels and climatic conditions, specific to the Indian agricultural context. Subsequently, these characteristics are sent into an "Optimized Artificial Neural Network (OANN)" that predicts fertilizer outcomes based on information already known. In particular, the weights of OANN are adjusted using back propagation and chain rule techniques to improve the classifier's prediction accuracy.