Inference for a partly linear autoregressive model with moving average errors

BIANCO, ANA; BOENTE, GRACIELA

Vigo

Simposio; International Seminar on Nonparametric Inference. (ISNI2008); 2008

IAP Attraction Pole, the Spanish Ministry of Science and Innovation, the Branch of Education and Universities of the Xunta de Galicia, the University of Vigo, IMS, ASA

In this talk, we generalize the partly linear autoregression model considered in the literature by including moving average errors when we want to allow a large dependence to the past observations. The strong ergodicity of the process is derived. A Fisher-consistent procedure to estimate the parametric and nonparametric components is provided together with a test statistic that allows to check the presence of a moving average component. Also, a Monte Carlo study is carried out to check the performance of the given proposals.