A Comparative Study of Extreme Learning Machine, Least Squares Support Vector Machine, Back Propagation Neural Network for Outlet Total Phosphorus Prediction

Abstract

Phosphorus is the main factor causing the degradation and eutrophication, which is important to strictly limit the emission of total phosphorus (TP) in wastewater treatment. Prediction outlet quality is an effective way to improve the monitoring level of sewage treatment. However, it is difficult to establish the mathematical model to predict outlet TP accurately because of the complex and changeable environment and the interrelated operation variables of sewage treatment. In this paper, three different data driven methods, extreme learning machine (ELM), least squares support vector machine (LS-SVM) regression method, and back propagation neural network (BPNN) are used for prediction of outlet TP concentration (TP(out)) base on observed data of inlet factors in sewage treatment plant, respectively. To get more accurate evaluation results of the ELM, LS-SVM and BPNN models, three quantitative standard statistical performance evaluation measures, mean square error (MSE), mean absolute percentage error (MAPE), and coefficient of determination (R2), are employed. The results show that the ELM method has remarkably superior performance on TP (out) prediction than peer models in outlet prediction of wastewater treatment.