GLC, MCMC and BAC: An Unified Approach?

Abstract

In this talk, I introduce two specific problems arising from GLC and MCMC. The first is that the asymptotic behavior of the maximum likelihood estimator in switch jump diffusion model, in which I proposed to capture the leptokurtic feature and volatility smile in financial time series. The second is that the stability property of dynamic weighting in Markov chain monte carlo simulation. A probability tool based on iterated random functions is, then, proposed to solve these two problems. Some possible further researches, include one in BAC, are also mentioned.