基于非线性回归最小二乘法的改进Gompertz模型参数估计

可靠性增长试验是产品研制阶段不可缺少的一项提高产品可靠性的工程项目。成败型试验的可靠性增长数据通常用Gompertz模型进行评估，但存在S形增长数据拟合精度差的不足。根据高斯一牛顿最小二乘法和多元线性回归最小二乘法原理，提出改进的Gompertz模型参数估计的非线性回归最小二乘法。用试验数据对改进后的模型进行验证，拟合度较好。

Parameter Estimation for Modified Gompertz Model Based on the Least-square-method of Nonlinear Regression

The reliabili .ty -growth testing is an inevitable project item for increasing the reliability of product in the process of developing products. The Gompertz model is usually used to estimate the data of the reliability - growth for success - and - failure test, but it shows poor fitness for S - shaped growth data. A least square method of nonlinear regression for modified Gompertz model parameter estimation is brought forward according to Gauss - Newton least square method and multiple linear regression least square method, and the fitness for the modified Gompertz model can be proved by using the data of test.