Deep Learning for Time Series Prediction in Fisheries Management

Abstract

The increasing popularity of deep learning has led to its widespread adoption in predicting future trends in time series data. This study investigates the applicability of Long Short-Term Memory (LSTM) and Gated Recurrent Units (G RU) models in forecasting demand within fisheries manage-ment. Specifically, a real-world case study focusing on scallop shell forecasting is presented. The dataset used encompasses comprehensive information on fish captures from January 1, 2015, to December 31, 2019. Through performance evaluation utilizing statistical metrics like Mean Squared Error (MSE) and Mean Absolute Error (MAE), our analysis reveals that the GRU-based approach outperforms LSTM in fisheries manage-ment applications. These findings underscore the potential of deep learning methodologies in enhancing demand forecasting accuracy and offer valuable insights for fisheries management practices.