Soft Measurement Model of Electrical and Mechanical Export Product Quality Based on Fuzzy Least Squares Support Vector Machines

Abstract

In terms of the fuzzy character of electrical and mechanical export product quality measurement system, self-adaptive variable metric chaos immune algorithm is used to optimize key parameters, and the soft measurement model of electrical and mechanical export product quality is established by using of Fuzzy Least Squares Support Vector Machines (FLS-SVM). The electrical and mechanical export product data during 1990 – 2010 show that the relative errors of warning estimated by the soft measurement model are less than 3.34%. Furthermore, factors which have major influence on soft measurements of electrical and mechanical export product quality are export product value, net trade volume, import product value and GDP successively.