Classification by Pairwise Coupling

By Trevor J. Hastie, Robert Tibshirani. 1997

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Auditing black-box models for indirect influence (2016, methodology): “We built two models that are notoriously opaque to simple examination; SVMs6 [12] and feedforward neural networks (FNNs).”
The offset tree for learning with partial labels (2008, methodology): “Multiclass regret on k classes is bounded by binary regret times k − 1 using the all-pairs reduction [7], which completes the proof.”
Sharing Visual Features for Multiclass and Multiview Object Detection (2007, background): “has size C C2 in which each column corresponds to a distinct pair of labels z1; z2; for this column, has þ1 in row z1, 1 in row z2, and 0 in all other rows; this corresponds to building all pairs of i versus j classifiers [15].”
Kernel-based methods for hyperspectral image classification (2005, background): “gies are available in the literature such as one-against-one [44], puncturing [45], or output-correcting-codes [46], which may re-”
A novel ensemble method for classifying imbalanced data (2015, methodology): “Finally the classification results of all the built classifiers are combined with the Hastie and Tibshirani's method [72].”
The fused Kolmogorov filter: A nonparametric model-free screening method (2014, methodology): “This strategy has been proven successful as a method for generalizing a binary classifier to its multiclass counterpart [Hastie and Tibshirani (1998)].”
To Combat Multi-Class Imbalanced Problems by Means of Over-Sampling Techniques (2016, methodology): “They used class decomposition schemes, one versus one (OAO) [29] and one versus all (OAA) [44], in order to apply the twoclass methods on multi-class imbalanced problems.”
Multicategory Support Vector Machines, Theory, and Application to the Classiﬁcation of Microarray Data and Satellite Radiance Data (2002, background): “See Zadrozny & Elkan (2002), Passerini, Pontil & Frasconi (2002), Price, Knerr, Personnaz & Dreyfus (1995) and Hastie & Tibshirani (1998).”
COVID-19 identification in chest X-ray images on flat and hierarchical classification scenarios (2020, unknown): “While the O-A-O technique, first used in [22], proposes to train a classifier for each possible pair of classes, ignoring the examples that do not belong to the related classes, the O-A-A approach, introduced by [23], builds a single classifier for each classe of the problem, considering the example”
Label ranking by learning pairwise preferences (2008, background): “…has been tried in the areas of statistics (Bradley and Terry, 1952; Friedman, 1996), neural networks (Knerr et al., 1990, 1992; Price et al., 1995; Lu and Ito, 1999), support vector machines (Schmidt and Gish, 1996; Hastie and Tibshirani, 1998; Kreßel, 1999; Hsu and Lin, 2002), and others.”
Multilabel classification via calibrated label ranking (2008, methodology): “Pairwise methods have previously been shown to improve performance for neural networks (Knerr et al. 1990, 1992; Price et al. 1995 ;L u and Ito1999), support vector machines (Schmidt and Gish 1996; Hastie and Tibshirani 1998 ; K resel1999; Hsu and Lin 2002), statistical learning (Bradley and Terry ”
A comparative study of feature selection and multiclass classification methods for tissue classification based on gene expression (2004, background): “One-versus-the-restmethod(Scholkopf and Smola, 2002; Bottou et al ., 1994), pairwise comparison (Kreeel, 1999; Hastie and Tibshirani, 1998; Friedman, 1996), error-correcting output coding (Dietterich and Bakiri, 1991, 1995; Allwein et al ., 2000) and multiclass objective functions (Weston and…”
Binarized Support Vector Machines (2010, background): “classes exist,C ={ 1,1}, since the multiclass case can be reduced to a series of two-class problems, as has been suggested e.g. in [ 20 , 22, 35].”
Automated annotation and classification of BI-RADS assessment from radiology reports (2017, unknown): “We trained the SVM classifier using sequential minimal optimization (SMO) [39,41,42] with a first-order polynomial kernel.”
Wavelet-based multiscale texture segmentation: Application to stromal compartment characterization on virtual slides (2010, methodology): “nary ones by decomposing it in one-against-one problems (Hastie and Tibshirani, 1998).”
Auditing Black-box Models by Obscuring Features (2016, methodology): “Since our goal is to examine the gradient feature auditing technique in the face of blackbox models, we built two models that are notoriously opaque to simple examination; SVMs6 [9] and Feedforward neural networks (FNNs).”