Institutional Repository
Institutional Repository Institutional Repository Login  Institutional Repository  
  • UOC Library |
  •  |
  •  |

Home >
Research >
Techonology and Communication >
Computer Science, Technology and Multimedia >
Parts of books or chapters of books >

Please use this identifier to cite or link to this item: http://hdl.handle.net/10609/1378
Title: A Hierarchical Approach for Multi-task Logistic Regression
Authors: Lapedriza Garcia, Àgata
Masip Rodó, David
Vitrià, Jordi
Issue Date: 2007
Type: Part or chapter of book
Citation: LAPEDRIZA, A.; MASIP, D.; VITRIÀ, J. (2007). "A Hierarchical Approach for Multi-task Logistic Regression". In: MARTÍ, J.; BENEDI, J.M.; MENDONÇA, A.M.; SERRAT, J. Lecture Notes in Computer Science. Springer. Núm. 4478. Pág. 258-265
Abstract: In the statistical pattern recognition eld the number of samples to train a classifer is usually insu cient. Nevertheless, it has been shown that some learning domains can be divided in a set of related tasks, that can be simultaneously trained sharing information among the different tasks. This methodology is known as the multi-task learning paradigm. In this paper we propose a multi-task probabilistic logistic regression model and develop a learning algorithm based in this framework, which can deal with the small sample size problem. Our experiments performed in two independent databases from the UCI and a multi-task face classification experiment show the improved accuracies of the multi-task learning approach with respect to the single task approach when using the same probabilistic model.
Description: Peer-reviewed
URI: http://hdl.handle.net/10609/1378
Appears in Collections:Parts of books or chapters of books

Scimago Journal Rank:

Add This:

RefWorks:

Ratings:
  0 (0 valuations)


Files in This Item:

File Description SizeFormat
Lapedriza_LNCS2007_Hierarchical.pdf131.7 kBAdobe PDFPreview  Download

Author names in Twitter


Author names in FriendFeed



Recommend this item

SFX Query

Items in Repository e are protected by copyright, with all rights reserved, unless otherwise indicated.

 

The library replies
A product of the Universitat Oberta de Catalunya Virtual Library