Interventions in ingarch processes

摘要: We study the problem of intervention effects generating various types outliers in a linear count time series model. This model belongs to class observation driven models and extends Gaussian within exponential family framework. Studies about covariates interventions for have largely fallen behind due fact that underlying process, whose behavior determines dynamics observed is not observed. suggest computationally feasible approach these problems, focusing especially on detection estimation sudden shifts outliers. To identify successfully such unusual events we employ maximum score tests, critical values finite samples are determined by parametric bootstrap. The usefulness proposed methods illustrated using simulated real data examples.