关键词: 离散数据, 数据拟合, 最小二乘支持向量机, 曲线拟合

Abstract: Least Squares Support Vector Machines（LS-SVM） is a novel support vector machines for discrete date fitting even with small samples．This paper approaches a data fitting model based on the LS-SVM．Motor characteristic curves are fitted with the intelligent model．The results show that the model has excellent learning ability and generalization ability，and can provide more accurate data fitting curve only with few observed samples compared with least square method．

Key words: discrete data, data fitting, Least Squares Support Vector Machines（LS-SVM）, curve fitting