Data clustering using moth-flame optimization algorithm

Abstract

In the presented work, the authors intend to detect and classify disease in cherry plants at a premature stage by analyzing its leaves. For experimental purposes, PlantVillage dataset has been used. Several machine learning models and a pre-trained CNN model have also been implemented for performance analysis. The performance analysis uses various metrics for evaluation like the number of epochs, AUC-ROC curve, recall, precision, and several other parameters. The proposed model when applied, the experimental results gave a better accuracy than the conventional ML algorithms. The implemented pre-trained model has achieved an approximate accuracy of about 99.89%.