Towards a Web-Based Fisheries Management System: An ANN Approach to Fish Weight Prediction

This research introduces a web-based fisheries management system that leverages Artificial Neural Networks (ANNs) to accurately predict fish weight. By utilizing a public dataset sourced from a fish market, encompassing seven species (Perch, Bream, Roach, Pike, Smelt, Parkki, and Whitefish), and employing advanced machine learning techniques, the developed system demonstrates exceptional performance in fish weight estimation. The dataset includes various fish characteristics such as length1, length2, length3, height, and width. The ANN model, trained on this comprehensive dataset, achieved an R-squared value of 0.987, indicating a strong correlation between predicted and actual fish weights. Furthermore, the model exhibited low mean squared error (MSE) of 1720.63 and root mean squared error (RMSE) of 41.48, demonstrating its accuracy and reliability. The system's ability to provide precise fish weight predictions is crucial for sustainable fisheries management, enabling informed decision-making regarding catch quotas, population assessments, and aquaculture practices. This innovative tool offers a significant advancement in fisheries management, promoting efficient resource utilization and conservation. Additionally, the system's user-friendly interface and intuitive design make it accessible to fisheries managers of all levels of expertise. As a comprehensive fisheries management information system, it offers a range of functionalities, including financial management, equipment rental, fish stock tracking, member interaction, and information dissemination.