The Statistical Inference Of Single Factor Interest Rate Model

With the constant improvement of China’s financial markets and deepening the degree of interest rate marketability, Modern term structure theory of interest rates plays an increasingly important role in a number of areas such as pricing the financial derivative products and risk management. Based on this view, this article chooses a single factor interest rate model as a starting point for research to explore parameter estimation methods of single factor interest rate model, then uses the generalized Bootstrap methods for statistical inference.After the establishment of single factor interest rate model, we need to estimate the model parameters. By using two different method compute transient density, we introduce two quasi MLE methods. They are Euler approximation and Perdersen method, because of the huge time consuming of Perdersen method, at last we choose the Euler approximation method.By comparing the goodness of the difference estimation methods, we have to compare the effectiveness, robustness, and other criteria of several methods.we introduced a new method named generalized bootstrap method to get the estimation of variance, then we create the confidence interval at last. The idea of generalized bootstrap method is clearly, we introduce some weights into to the likelihood function, then1/n substituted by this weight, as proved, with some regularity conditions, this method will get better result of variance estimation. Our simulation experiment proved this.In last chapter, we use Shibor data do some application. By setting the interest model using vasicek model, we get three parameters of this model. Then we use this model analyzed the yield curve of Shibor market.