[R] [Q] BIC as a goodness-of-fit stat

[Vincent Zoonekynd]

The following article gives some motivations for the BIC and
some rules to interpret a difference of BICs (if I recall correctly,
over 10: very strong difference, between 6 and 10: strong
difference, between 2 and 6: some difference):

  A.E. Raftery, Bayesian model selection in social research
  http://www.stat.washington.edu/tech.reports/bic.ps

Regards,

-- Vincent

On 06/03/06, Young-Jin Lee <youngjin.michael at gmail.com> wrote:
> Dear R-List
>
> I have a question about how to interpret BIC as a goodness-of-fit statistic.
> I was trying to use "EMclust" and other "mclust" library and found that BIC
> was used as a goodness-of-fit statistic.
> Although I know that smaller BIC indicates a better fit, it is not clear to
> me how good a fit is by reading a BIC number. Is there a standard way of
> interpreting a BIC value?
>
> Thanks in advance.
>
>         [[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
>