DEVELOPMENT OF MACHINE LEARNING BASED FISH SPECIES PREDICTION MODEL

Abstract: The correct identification and prediction of fish species within aquatic environment are important for effective fisheries management, biodiversity conservation, and ecosystem health assessment. Traditional methods of identifying the species of the fish species used to rely on the manual observation or invasive sampling techniques, which can be time consuming, labour intensive, and may not always provide real time data. In this research paper, we present with development of a machine learning based fish species prediction model with the use of the sample data we collected on the various fish species. Our research paper demonstrate the effectiveness of the machine learning based fish species prediction model using some machine learning based algorithm. The paper states the usefulness and effectiveness of the fish species prediction model developed through the use of machine learning algorithms over the traditional approaches used to identify the fish species. Our aim towards developing this model and following this approach was to help contribute to the people that use traditional approach in today in this modern world of technology. Our research is mainly towards the use of new technology to develop the fish species prediction model using CNN and transfer learning to increase the effectiveness of the model by using the layers of the already built predictive model and adding layers to this predictive model and making it more coustomized towards our approach of identifying the fish species.

Keywords: CNN(Convolutional Neural Network), transfer learning, machine learning.