Predictive Water Quality Monitoring in Aquaculture Using Machine Learning and IoT Automation

Abstract

This study presents an IoT and machine learning (ML)-driven intelligent aquaculture system for real-time water quality monitoring and automated decision-making. The system integrates IoT-enabled sensors to measure key parameters such as temperature, turbidity, dissolved oxygen, and water levels, with data processed by a Random Forest Classifier (RFC) for predictive analytics. The ML model classifies water quality conditions into optimal, warning, or critical states, triggering automated responses such as aeration control and water exchange. Experimental validation over six months demonstrates high classification accuracy (92.3%), improved automation, and reduced manual intervention, enhancing sustainability and efficiency in aquaculture management.