B0662
Title: Robustified Gaussian quasi-likelihood inference in YUIMA
Authors: Shoichi Eguchi - Osaka Institute of Technology (Japan) [presenting]
Abstract: Gaussian quasi-likelihood inference for stochastic differential equations is considered in the cases where the observations are obtained from the Levy process with the compound-Poisson jump and spike noise. For this problem, jumps and spike noises are regarded as outliers that disturb the parameter estimation and are constructed as an estimator without reference to the presence of the jump component and some spike noises, in addition to that of the drift term. The function which performs the estimation without reference to the presence of the jump component and some spike noises has been developed in the R package yuima. The estimation method is first overviewed and then the specification of the created function is explained. Some numerical examples are given in order to show how to use the function.