摘要
针对一般的最小二乘法忽略自变量的误差这一缺点,提出一种新的正交最小二乘法.该方法以正交距离残差平方和最小为衡量准则,并通过一定的算法剔除粗差和异常值,从而获得最佳的拟合曲线.算例结果表明稳健正交最小二乘拟合法结果更为可靠.
In traditional curve fitting methods, the error of independent variable was usually ignored. In the paper, the method of least squares orthogonal distance fitting is presented. This method gave a new criterion which is to minimize the square sum of orthogonal distance, and delete outliers of point clouds, so it is the best fitting curve. The results show that the proposed algorithm is reliable.
出处
《佳木斯大学学报（自然科学版）》
CAS
2013年第1期124-125,128,共3页
Journal of Jiamusi University：Natural Science Edition
基金
国家自然基金项目(51203061)
黑龙江省教育厅科学研究项目(11553114)
佳木斯大学科研项目(L2009-156)
关键词
最小二乘
曲线拟合
正交最小二乘
least square
curve fitting
least squares orthogonal distance fitting
