Given a log-likelihood, the number of observations and the number of estimated parameters, the average value of a chosen information criterion is computed. This facilitates comparison of models that are estimated with a different number of observations, e.g. due to different lags.

1 | ```
info.criterion(logl, n=NULL, k=NULL, method=c("sc", "aic", "aicc", "hq"))
``` |

`logl` |
numeric, the value of the log-likelihood |

`n` |
integer, number of observations |

`k` |
integer, number of parameters |

`method` |
character, either "sc" (default), "aic", "aicc" or "hq" |

Contrary to `AIC`

and `BIC`

, `info.criterion`

computes the average criterion value (i.e. division by the number of observations). This facilitates comparison of models that are estimated with a different number of observations, e.g. due to different lags.

a list with elements:

`method` |
type of information criterion |

`n` |
number of observations |

`k` |
number of parameters |

`value` |
the value on the information criterion |

Genaro Sucarrat, http://www.sucarrat.net/

H. Akaike (1974): 'A new look at the statistical model identification'. IEEE Transactions on Automatic Control 19, pp. 716-723

E. Hannan and B. Quinn (1979): 'The determination of the order of an autoregression'. Journal of the Royal Statistical Society B 41, pp. 190-195

C.M. Hurvich and C.-L. Tsai (1989): 'Regression and Time Series Model Selection in Small Samples'. Biometrika 76, pp. 297-307

G. Schwarz (1978): 'Estimating the dimension of a model'. The Annals of Statistics 6, pp. 461-464

Questions? Problems? Suggestions? Tweet to @rdrrHQ or email at ian@mutexlabs.com.

Please suggest features or report bugs with the GitHub issue tracker.

All documentation is copyright its authors; we didn't write any of that.