Hybrid Model Approach to Fish Price Prediction

The fishery industry plays an important role in national economic development, providing livelihoods and nutritional resources to millions. This makes the ability to predict fish prices a crucial advantage for fishermen, consumers and stockholders to secure and prepare for any unexpected or undesirable fluctuations in the same. However, the complexity of determining the price of fish due to various factors such as weight, species, freshness, season, and market conditions poses a challenge to both fishermen and consumers. This paper explores machine learning (ML) approaches for fish price prediction and proposes a hybrid framework that integrates variational mode decomposition (VMD), single spectrum analysis (SSA), improved beetle antennae search (IBES), long short-term memory (LSTM) and extreme gradient boost (XGBoost) to improve prediction accuracy. We employ data set-tuning techniques to minimize underfitting and overfitting, ensuring optimal model performance. The study focuses on the Konkan region, a crucial segment of India’s western coastline that includes the states of Goa and Maharashtra, analyzing a data set containing more than 130 species of seafood. Our system is designed for dual use, allowing fishermen to input catch data and customers to track price trends, providing both parties with information on market fluctuations. Additionally, the website serves as a tool for researchers and policymakers, enabling the collection and analysis of critical data. This contributes to broader studies on the socio-economic and dietary impacts of fish consumption, helping to understand the intricate economic trends within the fishery sector and its contribution to sustainable development. Through data visualization, users can grasp trends at a glance, making it an essential tool for economic decision-making in the fishery industry.