Document Detail


Hard or Soft Classification? Large-margin Unified Machines.
MedLine Citation:
PMID:  22162896     Owner:  NLM     Status:  Publisher    
Abstract/OtherAbstract:
Margin-based classifiers have been popular in both machine learning and statistics for classification problems. Among numerous classifiers, some are hard classifiers while some are soft ones. Soft classifiers explicitly estimate the class conditional probabilities and then perform classification based on estimated probabilities. In contrast, hard classifiers directly target on the classification decision boundary without producing the probability estimation. These two types of classifiers are based on different philosophies and each has its own merits. In this paper, we propose a novel family of large-margin classifiers, namely large-margin unified machines (LUMs), which covers a broad range of margin-based classifiers including both hard and soft ones. By offering a natural bridge from soft to hard classification, the LUM provides a unified algorithm to fit various classifiers and hence a convenient platform to compare hard and soft classification. Both theoretical consistency and numerical performance of LUMs are explored. Our numerical study sheds some light on the choice between hard and soft classifiers in various classification problems.
Authors:
Yufeng Liu; Hao Helen Zhang; Yichao Wu
Related Documents :
22025766 - Algorithms for reticulate networks of multiple phylogenetic trees.
23117306 - Practice on an augmented reality/haptic simulator and library of virtual brains improve...
22496816 - Patient-specific computational modeling of upper extremity arteriovenous fistula creati...
21839426 - Lateral thinking, from the hopfield model to cortical dynamics.
19859556 - Extreme cranial ontogeny in the upper cretaceous dinosaur pachycephalosaurus.
22410616 - A quasi-2-dimensional model for respiration of the cornea with soft contact lens wear.
Publication Detail:
Type:  JOURNAL ARTICLE    
Journal Detail:
Title:  Journal of the American Statistical Association     Volume:  106     ISSN:  0162-1459     ISO Abbreviation:  -     Publication Date:  2011 Mar 
Date Detail:
Created Date:  2011-12-13     Completed Date:  -     Revised Date:  -    
Medline Journal Info:
Nlm Unique ID:  01510020R     Medline TA:  J Am Stat Assoc     Country:  -    
Other Details:
Languages:  ENG     Pagination:  166-177     Citation Subset:  -    
Affiliation:
Department of Statistics and Operations Research, Carolina Center for Genome Sciences, University of North Carolina, Chapel Hill, NC 27599.
Export Citation:
APA/MLA Format     Download EndNote     Download BibTex
MeSH Terms
Descriptor/Qualifier:
Grant Support
ID/Acronym/Agency:
P01 CA142538-01//NCI NIH HHS; R01 CA085848-11//NCI NIH HHS; R01 CA149569-02//NCI NIH HHS

From MEDLINE®/PubMed®, a database of the U.S. National Library of Medicine


Previous Document:  Remote sensing of ambient particles in Delhi and its environs: estimation and validation.
Next Document:  Role of structural anisotropy of biological tissues in poroelastic wave propagation.