摘要
三因子模型是收益率曲线研究的一种经典方法。本文根据收益率曲线的几何特征,提出了收益率曲线三因子模型的一个直观定义;从四个维度对直观定义方法与经典三因子模型进行了实证比较,结果显示直观定义方法均表现较优,对收益率曲线的刻画更加精细;阐述了收益率曲线三因子的影响因素及三因子之间的非线性关系,并运用支持向量机(SVM)技术对未来利率走势进行预测,结果显示预测结果优于传统线性模型。
The three-factor model is a classical method for yield curve studies.Based on the geometric features of the yield curve,we propose a new approach—an intuitional definition of the three-factor model of the yield curve.Comparing the new approach with the classical three-factor model empirically from four perspectives,we find that the intuitional definition approach performs better invariably and it gives a more meticulous depiction of the yield curve.Besides,we study the influencing elements for the three factors of the yield curve,and find out the non-linear relations between the factors.Furthermore,support vector machine(SVM)is applied to yield curve prediction,and the results show that it performs better than traditional liner models.
作者
李晓花
胡志浩
Li Xiaohua;Hu Zhihao
出处
《中国社会科学院研究生院学报》
CSSCI
北大核心
2020年第5期90-108,共19页
Journal of Graduate School of Chinese Academy of Social Sciences
关键词
收益率曲线
三因子模型
直观定义
收益率曲线预测
yield curve
three-factor model
intuitional definition
yield curve prediction