@InProceedings{	  frank.ea:minimum:2011,
  abstract	= {The goal of model-order selection is to select a model
		  variant that generalizes best from training data to unseen
		  test data. In unsupervised learning without any labels, the
		  computation of the generalization error of a solution poses
		  a conceptual problem which we address in this paper. We
		  formulate the principle of ``minimum transfer costs'' for
		  model-order selection. This principle renders the concept
		  of cross-validation applicable to unsupervised learning
		  problems. As a substitute for labels, we introduce a
		  mapping between objects of the training set to objects
		  ofthe test set enabling the transfer of training solutions.
		  Our method is explained and investigated by applying it to
		  well-known problems suchas singular-value decomposition,
		  correlation clustering, Gaussian mixture-models, and
		  $k$-means clustering. Our principle finds the optimal model
		  complexity in controlled experiments and in real-world
		  problems such as image denoising, role mining and detection
		  of misconfigurations in access-control data.},
  author	= {Mario Frank and Morteza and Haghir Chehreghani and Joachim
		  M. Buhmann},
  booktitle	= {ECML PKDD 2011: European Conference on Machine Learning
		  and Principles and Practice of Knowledge Discovery in
		  Databases},
  copyright	= {Springer},
  copyrighturl	= {www.springerlink.com},
  language	= {USenglish},
  month		= {Sept},
  pdf		= {papers/2011/ECML2011_TransCosts.pdf},
  publisher	= {Springer},
  title		= {The Minimum Transfer Cost Principle for Model-Order
		  Selection},
  year		= 2011,
  user		= {mafrank}
}



Reference Type: Book Chapter
Author: Frank, Mario
Author: Chehreghani, Morteza
Author: Buhmann, Joachim
Editor: Gunopulos, Dimitrios
Editor: Hofmann, Thomas
Editor: Malerba, Donato
Editor: Vazirgiannis, Michalis
Primary Title: The Minimum Transfer Cost Principle for Model-Order Selection
Book Title: Machine Learning and Knowledge Discovery in Databases
Book Series Title: Lecture Notes in Computer Science
Copyright: 2011
Publisher: Springer Berlin / Heidelberg
Isbn: 978-3-642-23779-9
Start Page: 423
End Page: 438
Volume: 6911
Url: http://dx.doi.org/10.1007/978-3-642-23780-5_37
Doi: 10.1007/978-3-642-23780-5_37