论文部分内容阅读
阻尼最小二乘法或称Ｌｅｖｅｎｂｅｒｇ—Ｍａｒｑｕｒｔ法（简称Ｌ─Ｍ法）是我们在时间序列分析中经常用到的一种实用的非线性最小二乘算法。本文指出了常用阻尼最小二乘算法的不足之处，并提出了一种实用的改进算法─—三点搜索法。实例表明改进的算法不仅可以提高计算速度，而且可以得到更好的求解结果。
Damping Least Squares or the Levenberg-Marqurt method (L ─ M method) is a practical nonlinear least squares algorithm that we often used in time series analysis. This paper points out the shortcomings of the commonly used least squares algorithm and proposes a practical improved algorithm ─ ─ three-point search method. The examples show that the improved algorithm can not only improve the computational speed, but also get a better solution.