论文部分内容阅读
We consider the empirical likelihood estimation problem for the process of OrnsteinUhlenbeck type with discretc observations.The approach is based on the conditional characteristic function.The maximum empirical likelihood estimator is proved to be consistent and asymptotically normal.Moreover, this estimator is asymptoti cally efficient under mild conditions.When the background driving Levy process is a Levy process of type A or B, thc intensity parameter can be exactly recov ered, and the maximum empirical likelihood estimator with the intensity parameter plugged in is studied.Testing procedures based on empirical likelihood ratio statis tic are developed for parameters and for estimation equations respectively.Monte Carlo simulations are also designed to evaluate the performance of maximum em pirical likelihood estimator and empirical likelihood ratio testing.[Joint work with Shuguang Sun]