Universitas Dian Nuswantoro conducted research on optimizing parameter combinations in Convolutional Neural Networks for classifying nutrient deficiencies in rice plants. The activity involved Sari, Yuslena; Pramunendar, Ricardus Anggi; Mambang, M.; Cipta, Subhan Panji; Alzami, Farrikh; and Sunardi, Sunardi as research participants. The research addressed challenges in precision agriculture associated with population growth, shrinking arable land, and climate change. It focused on efficient plant nutrient management to enhance productivity while supporting sustainability.
The research introduced an Improved Grey Wolf Optimizer method for determining suitable Convolutional Neural Network parameter values for detecting nutrient deficiencies. The parameters included the number of epochs, dropout rate, hidden layer size, learning rate, activation function, and optimizer. The Improved Grey Wolf Optimizer incorporated an adaptive parameter adjustment mechanism to improve the balance between exploration and exploitation, reduce the risk of premature convergence, and improve computational efficiency. The research used the Nutrient Deficiency Symptoms in Rice dataset from Kaggle.
The activity compared the Improved Grey Wolf Optimizer with the original Grey Wolf Optimizer, Particle Swarm Optimization, Whale Optimization Algorithm, hybrid Grey Wolf Optimizer and Whale Optimization Algorithm, and the FOX optimizer. The resulting parameters were 8 epochs, a dropout rate of 0.1, a hidden layer size of 128, a learning rate of 0.002, the ReLU activation function, and the Adam optimizer. The method achieved a training accuracy of 86.77%, a test accuracy of 82.43%, and a training time of 109.50 seconds. The original Grey Wolf Optimizer and Whale Optimization Algorithm showed slower convergence and lower accuracy. Particle Swarm Optimization incurred higher computational costs, while the FOX optimizer and hybrid method were less efficient and stable.
This UDINUS activity relates to the stated goals through research on rice nutrient deficiency detection, precision agriculture, climate change context, and sustainable crop management. It provides an optimization-based approach for using Convolutional Neural Networks to detect nutrient deficiencies in rice plants.