Evaluation of ML Models for Detection and Prediction of Fish Diseases: A Case Study on Epizootic Ulcerative Syndrome

Abstract

Fish diseases pose significant threats to the aquaculture industry, leading to adverse economic and environmental impacts. The early detection and diagnosis of fish diseases are crucial for effective disease control and prevention. In this study, fish images with Epizootic Ulcerative Syndrome (EUS) disease symptoms were collected along with non-infected fish images, which were then augmented to obtain a larger dataset. The pre-processed images were analysed using different machine learning algorithms, including “decision tree, logistic regression, naive Bayes, support vector machine (SVM), and multi-layer perceptron (MLP)”. The study found that the SVM-based system was effective in detecting EUS disease in fish, achieving an accuracy of 85.24% on the original dataset using a polynomial kernel, and 82.75% on the augmented dataset using a Gaussian kernel. These results suggest that SVM-based systems can be used for the early detection and prevention of EUS disease in fish, highlighting their potential for application in the aquaculture industry. Furthermore, the study indicates the importance of dataset augmentation in improving the accuracy of machine learning models in detecting fish diseases. The findings of this study can serve as a foundation for future research on the development of effective machine learning models for the early detection and diagnosis of various fish diseases.