The rapid industrialization and urbanization of the Uruan Local Government Area of Akwa Ibom State, Nigeria, have significantly contributed to heavy metal contamination in the Nwaniba River. This research assesses the ecological risks associated with contamination using advanced machine learning techniques. Water, sediment, and marine organism samples were analyzed using ICP-OES to quantify seven heavy metals: iron (Fe), copper (Cu), nickel (Ni), lead (Pb), zinc (Zn), chromium (Cr), and manganese (Mn). Three machine learning models—Random Forest (RF), Support Vector Machines (SVM), and Artificial Neural Networks (ANN)—were employed to predict heavy metal concentrations and evaluate ecological risks. Random Forest exhibited the highest predictive accuracy (R² = 0.88), followed by ANN (R² = 0.83) and SVM (R² = 0.52). Results indicated significant contamination levels in sediments and bioaccumulation in marine organisms, particularly for Cu, Zn, and Mn, posing risks to aquatic life and public health. Although water quality parameters such as dissolved oxygen and pH generally met WHO, EPA, and NSDWQ standards, the high levels of heavy metals in sediments and organisms underscore the need for continuous monitoring and remediation. The study highlights the critical role of machine learning in enhancing ecological risk assessment and informs strategies for the sustainable management of aquatic ecosystems.
