摘要

This paper develops a discrete time microstructure jump model with nonhomogeneous Poisson process. Because the jump intensity is changed with time, the jumps are detects using a new nonparametric method proposed by Suzanne S. Lee. On the basis of this analysis, its parameters are estimated by combination of extended Kalman filter(EKF) and maximum likelihood (ML) method. Simulation results and case study show the effectiveness of the estimation method.

全文