Cholesky Decomposition for the Vasicek Interest Rate Model

摘要

This paper concerns the estimation of parameters in the ``Vasicek Interest Rate'' model under a Bayesian framework. These popular models are challenging to fit with Markov chain Monte Carlo (McMC) methods as the structure of the model leads to considerable autocorrelation in the chains. Accordingly, we demonstrate that a simple re-parameterisation using the Cholesky decomposition can greatly improves the performance of the McMC algorithm and hence lead to valid Bayesian inference on the Vasicek model.