Data-Driven Predictions of Fish Production: Applying Regression Methods in Aquaculture

Abstract

In the context of the industrial revolution 4.0, the fisheries sector is experiencing significant technological advancements, enhancing the effectiveness and efficiency of aquaculture in Indonesia, which plays a crucial role in food security. Aquaculture, defined in various ways, includes the cultivation of fish in diverse environments such as fields, rice paddies, and marine settings, and is seen as a key driver of economic growth by the Ministry of Marine Affairs and Fisheries (KKP), particularly through new initiatives that focus on export-based fisheries cultivation and the development of fisheries villages rooted in local wisdom. With untapped potential in aquaculture land that can generate substantial profit margins and support industrial development, it remains a sector that can benefit all levels of society. However, sustainable practices must be prioritized due to declining fish populations caused by overfishing, necessitating restocking efforts driven by accurate demand forecasting and careful consideration of market fluctuations affecting selling prices. This study introduces a predictive system aimed at improving harvest partners' profitability by employing various regression methods, including multiple linear regression, support vector regression, polynomial regression, and random forest regression, to analyze data from aquaculture activities that include key factors such as seeding information, feeding monitoring, mortality rates, and average fish weight. The process involves meticulous data preparation, visualization, and establishment of a training dataset, followed by rigorous validation and precision testing of predictive models using metrics such as mean absolute percentage error (MAPE) and root mean square error (RMSE), ultimately providing a robust framework for improving decision-making in fisheries cultivation.