On full order autoregressive model fitting

Abstract

This is a study of the relative performance of nine automatic autoregressive order determination criteria for full-order modeling using the least squares approach. Of the nine: AIC, FPE, CAT₂, S, FPE4, Φ, SIC, BIC and CAT₃, we have found the trio, AIC, FPE and CAT₃, to be the most consistent. The rest underestimate relatively.

Journal of the Nigerian Association of Mathematical Physics, Volume 15 (November, 2009), pp 117 – 122