Evolutionary Fuzzy Function with Support Vector Regression for the Prediction of Concrete Compressive Strength

Abstract

The main purpose of this paper is to develop an evolutionary fuzzy function with support vector regression (EFF-SVR) model to predict the compressive strength of concrete. Fuzzy functions alter conventional fuzzy system modelling methods structurally. They take advantage of utilizing membership values calculated by fuzzy c-mean (FCM) clustering, and their possible transformations, as additional explanatory variables augmented to the original input space. Since support vector regression (SVR) methods have considerable capability of minimizing both empirical and complexity risks simultaneously, the hybrid model of EFF-SVR is expected to yield robust results. Finally, the generalization capability and robustness of EFF-SVR are compared with some existing system modelling methods, i.e., artificial neural network (ANN), adaptive neural-fuzzy inference system (ANFIS), fuzzy function with least squared estimation (FF-LSE), and improved FF-LSE. The results show that EFF-SVR has a great ability as a feasible tool for prediction of the concrete compressive strength.