我国国债收益率特征研究——基于国债利率期限结构与收益率关系的实证分析

Research on Characteristics of National Debt Yield in China——An Empirical Research on Relationship Between Term Structure and Treasury Yields

本文针对遗传算法的多项式样条函数利率期限模型中,单纯遗传算法在进化过程中容易丧失样本多样性的缺陷,提出一种基于成长机制的遗传算法,改进单纯遗传算法的求解能力,并利用2016年10月21日国债市场上10只性质一致的国债进行实证比较。实验证明,从定价误差角度来看,所提算法优于后者,能够较为精确地描述国债收益率与期限之间的关系。最后为我国国债市场的发展和相关政策制定提供相应参考建议。

Directed at the simple genetic algorithm that is easy to lose the sample diversity of evolution process in interest Rate term structure model of polynomial spline function based on genetic algorithms. The paper proposes a genetic algorithm based on growth mechanism, improves the solving ability of simple genetic algorithm, and uses ten national debt which are consistency property in October 21, 2016 to empirical comparison. Experiments shows that the proposed algorithm is superior to the latter, andcan describe the relationship between bond yield and maturity more accu-ratelyfrom the point of view of mispricing. Finally for the development and relevant policy-making ofnational debt market in China, provides reference suggestions.