中文摘要：

将遗传算法引入扩展Nelson-Siegel模型估计中,并将其用于国债收益率曲线的估计。实证分析表明,基于遗传算法的扩展Nelson-Siegel模型在收益率曲线拟合和估计方面明显优于基于三次样条插值的息票剥离法及基于非线性回归的扩展Nelson-Siegel模型。基于此,利用基于遗传算法的扩展Nelson-Siegel模型对所选取的三个样本交易日的收益率曲线进行估计和分析,发现金融危机中后期的收益率曲线较金融危机初期的收益率曲线有了显著变化,主要表现在收益率曲线整体水平下降,但不同部分下降的幅度不同,并且收益率曲线的斜率和曲度均明显增大。以上变化主要是金融危机背景下货币政策调整以及市场信心变化共同作用的结果。

英文摘要：

This paper introduces an estimation approach for the Extended Nelson-Siegel model of termstructure of interest rates based on the genetic algorithm（hereafter GA）.Then an empirical comparison ismade between the bootstrap method based on cubic spline interpolation,the Extended Nelson-Siegelmodel based on nonlinear regression and the model introduced in this paper,which shows that the last oneoffers the best performance of the three methods in yield curve fitting and estimation.So yield curves areestimated and analyzed based on the method proposed for the three sample dates and then it is found thatthere are significant differences between yield curves of the initial stage and mid-and late one of the ongo-ing financial crisis in three aspects：the level of the yield curve drops,with different magnitudes for differ-ent parts,while both the slope and curvature increase significantly,which is attributed to the shift of themonetary policy and the recovery of the market confidence in the context of the financial crisis.