摘要
曲线拟合是实验数据处理的基本方法之一。将曲线拟合方法归结为有理论模型和无理论模型两类,据此,对曲线拟合的一般思路和重要方法进行了讨论。对两类方法进行了比较,并将它们联合用于对材料流变状态的速率-微分型本构模型的曲线拟合。
Curve fitting is one of the basic methods in experimental data processing. In this paper, the methods of curve fitting are classified by existence of theoretic model. According to the above (classification), the paper discusses the general and important methods of curve fitting. Then it compares the two species of method and addresses a new method based on neural networks and least square. Finally, the new method is used for curve fitting of rheologic modelvelocity-differential constitutive relations.
出处
《成都理工大学学报（自然科学版）》
CAS
CSCD
北大核心
2004年第1期91-95,共5页
Journal of Chengdu University of Technology: Science & Technology Edition
基金
国家973项目(G1999022511)
关键词
数据处理
曲线拟合
最小二乘
神经网络:本掏关系
data processing
curve fitting
least square
neural networks
constitutive (relations)
