Modifications of BIC for data mining under sparsity

Florian Frommlet ()
Additional contact information
Florian Frommlet: Medical University Vienna

A chapter in Operations Research Proceedings 2011, 2012, pp 243-248 from Springer

Abstract: Abstract In many research areas today the number of features for which data is collected is much larger than the sample size based on which inference is made. This is especially true for applications in bioinformatics, but the theory presented here is of general interest in any data mining context, where the number of “interesting” features is expected to be small. In particular mBIC, mBIC1 and mBIC2 are discussed, three modifications of the Bayesian information criterion BIC which in case of an orthogonal designs control the family wise error (mBIC) and the false discovery rate (mBIC1, mBIC2), respectively. In a brief simulation study the performance of these criteria is illustrated for orthogonal and non-orthogonal regression matrices.

Keywords: False Discovery Rate; Bayesian Information Criterion; Orthogonal Design; Hadamard Matrice; Multiple Interval Mapping (search for similar items in EconPapers)
Date: 2012
References: Add references at CitEc
Citations:

There are no downloads for this item, see the EconPapers FAQ for hints about obtaining it.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:spr:oprchp:978-3-642-29210-1_39

Ordering information: This item can be ordered from
http://www.springer.com/9783642292101

DOI: 10.1007/978-3-642-29210-1_39

Access Statistics for this chapter

More chapters in Operations Research Proceedings from Springer
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().