On the Selection of Subset Autoregressive Time Series Models

Abstract

The estimation of subset autoregressive time series models has been a difficult problem because of the large number of possible alternative models involved. However, with the advent of model selection criteria based on the maximum likelihood, subset model fitting has become feasible. Using an efficient technique for evaluating the residual variance of all possible subset models, a method is proposed for the fitting of subset autoregressive models. The application of the method is illustrated by means of real and simulated data.

U. S. Influences on Foreign Monetary Policy

“…Second, I employ Haggan and Oyetunji's (1984) subset procedure for choosing lags. Researchers commonly estimate equations like Yt °t° + °tlYt-1 + * * * 0tTYt-T + lWt-1 + * * * + SWt-S * (9) Searching for the appropriate lags generally is limited to using one of the above statistical criteria to determine the maximum order of T and S. Haggan and Oyetunji (1984) contend that researchers also should examine subsets of the lags, that is, subset time series models.…”

The export-productivity relationship: A time series representation for Austria

“…The resulting model, which is unlikely to be too far from the best out of approximately 284 (the total number of subsets), will be denoted as SMAR (for subset model AR). Although the imposition of zero restrictions is sometimes criticized (see, e. g., Judge --Brock, 1983), there is now growing empirical evidence for subset models outperforming unrestricted vector AR (Haggan --Oyetunji, 1984, Kunst --Neusser, 1986). …”

“…Variable (subset) selection started from a maximal model obtained by setting the maximal lag of the ~ii to the one identified by univariate AIC and of the off-diagonals to 8. Although Haggan --Oyetunji (1984) see no difficulties with regard to the computer time involved with computing the AIC (or some other criteria) of all possible subset models in a univariate framework, we faced such problems and rather used naive selection and elimination based on consecutive omission of the least significant regressor within each of the single equations and calculating the resulting AIC values. This selection procedure, however, might have reduced the model too far.…”

Subset ARMA Model Identification Using Genetic Algorithms

“…Recentlly, a class of subset autoregressive (SAR) models has been proposed by several reseachers, such as McClave (1975), Penm and Terrell (1982), Haggan and Oyetunji (1984), Yu and Lin (1991) and Zhang and Terrell (1997).…”