Eutrophication Prediction Model of Bohai Bay Based on Optimized Support Vector Machine

Abstract

In this research, optimized SVM models were designed to describe eutrophication processes, based on the field measured data from Bohai Bay. A new data-driven model called Support Vector Machine (SVM) based on structural risk minimization principle was presented, which minimized a bound on a generalized risk. In the eutrophication model, the Principal Component Analysis (PCA) was used to identify the model inputs. After data scaling, cross-validation via parallel grid search and genetic algorithm were respectively employed to select the optimal parameters of SVM. The model performance was evaluated by means of the squared correlation coefficient R <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> and the Root Mean Square Error (RMSE). The results suggest that parameters optimization is very important and necessary for SVM, and SVM-GA (Genetic Algorithm integrated with SVM) possesses slightly better searching optimization ability. It was shown that this optimized SVM techniques could be applied to predict the concentration of Chlorophyll_a in Bohai Bay and capture the non-linear information in eutrophication processes.