一类纯非线性回归模型的新解法及其应用  

New Solution of a Kind of Pure Nonlinear RegressionModel and Its Application

作　　者：张子贤[1] 路梅[1] 

机构地区：[1]河北工程技术高等专科学校水利工程系,河北沧州061001

出　　处：《水电能源科学》2002年第4期57-59,共3页Water Resources and Power

摘　　要：针对只有一个参数是以非线性形式出现的纯非线性回归模型,提出一种新方法确定模型中的回归系数。实例表明,新方法的拟合精度与Gauss-Newton法相当,而计算要比Gauss-Newton法简单、方便。Applying the principle of least square estimation and numerical solution of nonlinear equation, a new method is used to define regression coefficients in the pure nonlinear regression model that only one parameter appears in the nonlinear form. Example shows that fit accuracy of this method is as good as GaussNewton method, but calculation is simpler and easier than GaussNewton method.

关 键 词：纯非线性模型 最小二乘法 回归系数 非线性方程 数值解法 拟合精度 

分 类 号：O212.1[理学—概率论与数理统计] P333.[理学—数学]