利率互换定价主要是受到金融市场利率期限结构的影响。依据美国国库券收益率和互换利率数据,结合我国国库券收益率数据,采用非参数的支持向量机预测模型模拟出利率期限结构;在已知利率期限结构的基础之上,采用支持向量机的方法模拟估计出利率互换的固定利率,从而构造出一种系统的利率互换定价方法。通过实证检验,结果表明基于支持向量机的定价方法是可行的,且精确度也比其他定价方法要高。
The pricing of interest rate swap is mainly influenced by term structure of interest rate in money market. In the light of the data of American treasury bill earning rate and interest rate swap, in combination of our country's data of treasury bill earning rate, the term structure of interest rate is simulated by using nonparametric support vector machine forecasting model, and on the basis of this, a systemic method of interest rate swap pricing is formulated by using support vector machine method to estimate fixed interest rate of swap. Through the analysis of demonstration, the result suggests that this pricing method based on support vector machine neural network is effective and feasible, and it has higher precision than other methods.