Machine learning, Water quality -- Forecasting, Support vector machines, Supervised learning (Machine learning)
The objective of this research is to explore prediction accuracy of water quality factors, with techniques and algorithms in machine learning consisting of a variation of support vector machines - Support Vector Regression (SVR) and the gradient boosting algorithm Extreme Gradient Boosting (XGBoost). Both the XGBoost and SVR algorithms were used to predict nine different factors with success rates ranging from 79% to 99%. Parameters of these algorithms were also explored to test the prediction accuracy levels of individual water quality factors. These parameters included normalizing the data, filling missing data points, and training and testing on a large set of data.
Joslyn, Kathleen, "Water Quality Factor Prediction Using Supervised Machine Learning" (2018). REU Final Reports. 6.