Performance Comparison of Two Methods used for LMCS Data Fusion

This paper presents the basic structure and algorithm of two networks: multilayer perceptron (MLP) and support vector machines (SVMs). Both networks are applied to lumber moisture content sensor (LMCS) data fusion to solve the nonlinearity caused by ambient temperature and other disturbance factors. The result of comparative analysis is given. The training algorithm of MLP may be trapped in a local minimum and has a difficult task to determine the best architecture. SVM based on statistical learning theory and structural risk minimization is proposed to deal with these problems. It shows that solutions obtained by SVM training seem to be more robust and better generalization performance compared to MLP training