Accurate Dissolved Oxygen Prediction for Aquaculture Using Stacked Ensemble Machine Learning Model

Dissolved oxygen (DO) is the most vital water quality parameter that directly indicates the survival of aquatic life. Therefore, accurate DO prediction is essential for aquaculture water quality management for sustainable and profitable aquaculture production. Machine learning (ML) models have been successfully employed for water quality prediction. However, DO undergoes dynamic changes, which are nonlinear and complex, making accurate prediction of DO using conventional statistical methods and ML models a challenging task. To resolve this in this work, we propose a stacked ensemble ML model combining three different ML models as base learners and one ML model as a meta-learner to improve the DO prediction accuracy. The effectiveness of the stacked ensemble ML model has been evaluated using two different water quality datasets. The experimental results show that the stacked ensemble ML model achieves significant accuracy improvement compared with standalone ML models.