题 目：Estimation and Testing for Markov Processes via Conditional Characteristic Function

汇报人：Prof.Song Xi Chen（Department of Statistics，Iowa State University）

时 间：2008年1月8日（周二）上午10：00-11：00

地 点：J9集团国际站120室

摘 要：Markov processes encompasses a large class of models which are used in financial and economical modeling these days.For many of these models, the transitional distribution of the process is not available for a conventional statistical/econometrical inference via the maximum likelihood. We propose a nonparametric likelihood ratio for parameters of a Markov process via its conditional characteristic function, whose form are more likely to be available. The likelihood can be used for parameter estimation and model specification testing of a parametric Markov process which can be a L\'evy driven process. The proposed inference framework can be applied to a range of sampling frequency of the process including high frequency data. We quantify the estimation efficiency and power of the model specification test asymptotically. Results from numerical simulations are reported for diffusion, jump diffusion and L\'evy processes