Predictive Conservation: Applying Machine Learning to Identify Overfishing Risks Linked to Water Quality and Temperature

Overfishing continues to threaten marine biodiversity and the livelihoods of communities reliant on fishing industries. This paper proposes a machine learning–driven framework to predict overfishing risks based on key environmental factors, notably water quality and temperature. By analyzing a comprehensive dataset comprising species attributes and regional variables, we compare the performance of four models—Random Forest, Logistic Regression, Support Vector Machine, and Neural Network—in identifying potential overfishing hotspots. The results illustrate the capacity of these algorithms to reveal critical indicators (e.g., fish population, average size, water pollution levels) that influence the likelihood of stock depletion. Beyond ecological considerations, we highlight how data-driven analyses can inform business decisions and policy interventions, guiding sustainable harvesting strategies and fostering resilience against environmental volatility. The insights presented herein underscore the importance of integrating predictive modeling into conservation practices, offering a proactive approach to fishery management for both ecological preservation and economic stability.