Prediction Method for Machining Quality Based on Weighted Least Squares Support Vector Machine

Abstract

A new machining error prediction approach, which is based on the weighted least squares support vector machine (LS-SVM), was given. The nearer sample was set a larger weight, while the farther was set the smaller weight in the history data. In the same condition, the results show that the prediction accuracy of the weighted LS-SVM is 40% higher than that of the standard LS-SVM. Compared with other more modeling approaches, the prediction effect indicates that the proposed method is more accurate and can be realized more easily. It provides a better way for on-line quality monitoring and controlling of dynamic machining