Price Forecasting of a Small Pelagic Species in a South American Supply Center: A Machine Learning Approach

Small pelagic fish play a key role in human nutrition, especially in emerging countries, as they are affordable protein sources and provide income for fishing communities. Despite their nutritional benefits for human health, prices are the main factor when choosing seafood as diet components, which highlights the relevance of an economic analysis, since changes in fish prices might alter the feeding patterns of populations worldwide. This study analyzed the price dynamics of the Sardinella brasiliensis in one of the main markets in northeast Brazil and employed machine learning techniques to forecast future prices. The dataset was obtained from the Pernambuco Supply and Logistics Center website, and it was modeled using the FbProphet library in addition to a long short-term memory (LSTM) neural network in order forecast future prices. Both algorithms reached low error metrics, but LSTM performed significantly better, showing its usability in the economic context of marine products.