AIC FOR ERGODIC DIFFUSION PROCESSES
FROM DISCRETE OBSERVATIONS


Masayuki UCHIDA
Kyushu University
Fukuoka, Japan

Nakahiro YOSHIDA
University of Tokyo
Tokyo, Japan


ABSTRACT

For discretely observed ergodic diffusion models, Akaike's information criterion based on the expected Kullback-Leibler information is considered. The information criterion is applied to model selection problems of ergodic diffusion models. Asymptotic properties of the information criterion are also discussed.



Back to the Workshop's Program