Simulation of Methyl Tertiary Butyl Ether Concentrations in River-Reservoir Systems Using Support Vector Regression

Abstract

Mathematical and numerical models are used to simulate the transport of pollutants released into a water body. Such simulations can be computationally burdensome, however. One approach to overcome computational burdens associated with the simulation of pollutant transport is to use data-mining tools. The aim of this study is to simulate the concentration of methyl tertiary butyl ether (MTBE) at various locations within a river-reservoir system using the support vector regression (SVR) data-mining tool. The SVR tool is optimized by means of a genetic algorithm (GA). This paper’s results indicate that the developed and optimized SVR tool is more accurate than artificial neural networks (ANN) and genetic programming (GP) when judged by the correlation coefficient of regression analysis (R2).