bestModels {artfima} | R Documentation |

## Best BIC Models

### Description

ARIMA(p,0,q), ARFIMA(p,0,q) and ARTFIMA(p,0,q) models are fit for various p=0,1,..., and q=0,1,... and the best models according to the BIC criterion are selected.

### Usage

```
bestModels(z, parMax = 4, nbest = 4, likAlg = c("exact", "Whittle"),
d=0, ...)
```

### Arguments

`z` |
time series data |

`parMax` |
maximum number of parameters - see Details |

`nbest` |
number of models in selection |

`likAlg` |
likelihood method to use |

`d` |
regular differencing parameter indicating the number of times to difference |

`...` |
optional arguments for artfima such as lambdaMax |

### Details

`numPar = K`

, where K is
the number of structural models defined by
`K = p+q+n(glp)`

,
where `n(glp) = 0, 1, 2`

according as the model is ARIMA, ARFIMA or
ARTFIMA respectively.

These models are ranked according to the AIC/BIC criterion and the best ones are shown.

The plausibility is shown. This is defined for AIC by the eqn
`p(AIC) = exp(0.5*(min(AIC)-AIC))`

,
where AIC is the vector of AIC values. Similarly for the BIC.

### Value

An S3 list object, "bestmodels". Output is provided using the print method for the "bestmodels"

### Note

There are often small differences in the likelihood among a group of 5 or more of the best models. So the "exact" and "Whittle" likelihood methods may produce a different ranking of the models. For this reason the "exact" likelihood method may be preferred.

### Author(s)

A.I. McLeod

### See Also

`best_glp_models`

`print.bestmodels`

### Examples

```
## Not run:
data(ogden)
\dontrun{ #about 10 seconds
bestModels(ogden)
}
## End(Not run)
```

*artfima*version 1.5 Index]