Predict water quality based on multiple kernel least squares support vector regression and genetic algorithm

A multiple kernel least squares support vector regression model for prediction of water quality is put forward. In this model, samples of water quality are mapped into a high dimensional feature space by a committee nonlinear function, and then fitted with a linear regression model. The regularization parameters and the kernel parameters of the model are optimized based on genetic algorithm. The efficacy of the model is demonstrated on the East river, and the result shows that the proposed approach has a better prediction accuracy and generalization performance than the least squares support vector regression with a single kernel.