The weighted average information criterion for multivariate regression model selection

Tiee Jian Wu, Pinyuen Chen, Yanjun Yan

Research output: Contribution to journalArticlepeer-review

9 Scopus citations


We propose a consistent criterion WIC vc (vector corrected weighed average information criterion) for model order selection in multivariate linear regression models. The WIC vc is a weighted average of the asymptotically efficient criterion KIC vc (vector corrected Kullback information criterion) and the consistent criterion MBIC (multivariate Bayesian information criterion). The WIC vc behaves like KIC vc in small samples and behaves like MBIC in large samples. A numerical study comparing the performance of the proposed criterion with several available model selection criteria has been done. It shows that, over a wide range of small, moderate and large sample sizes, the WIC vc is more stable in comparison to other criteria in the study; that is, the WIC vc is either as good or comes in a strong second, whereas other criteria vary more in performance ranking. Therefore, the WIC vc is a very reliable and practical criterion.

Original languageEnglish (US)
Pages (from-to)49-55
Number of pages7
JournalSignal Processing
Issue number1
StatePublished - Jan 2013


  • AIC
  • AIC
  • BIC
  • Consistent criterion
  • KIC
  • Order selection

ASJC Scopus subject areas

  • Control and Systems Engineering
  • Software
  • Signal Processing
  • Computer Vision and Pattern Recognition
  • Electrical and Electronic Engineering


Dive into the research topics of 'The weighted average information criterion for multivariate regression model selection'. Together they form a unique fingerprint.

Cite this