A comparison of methods for multiclass support vector machines

By Chih-Wei Hsu, Chih-Jen Lin. IEEE Trans. Neural Networks, 2002

6,877 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/7f755d620b57acf27a16ff95923c5677ff8198bb

How researchers see this paper

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

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

Are sparse representations really relevant for image classification? (2011, methodology): “We were able to achieve our best results on the CIFAR-10 datas et by solving an expensive multi-class bound-constrained SVC problem with γ = 8 using the BSVM software [1].”
Tuning support vector machines for biomedical named entity recognition (2002, methodology): “Among several methods for constructing a multi-class SVM (Hsu and Lin, 2002), we use a pairwise method proposed by Kreßel (1998) instead of the one-vs-rest method used in (Yamada et al.”
Deep Learning using Linear Support Vector Machines (2013, background): “Hsu & Lin (2002) discusses other alternative multiclass SVM approaches, but we leave those to future work.”
Ordinal Regression Methods: Survey and Experimental Study (2016, background): “CSSVC Cost-Sensitive Support Vector Classifier (CSSVC) [59]”
Label ranking by learning pairwise preferences (2008, methodology): “Pairwise classification has been tried in the areas of statistics [9, 23], neural networks [44, 45, 55, 48], support vector machines [58, 32, 46, 35], and others.”
Continuous body and hand gesture recognition for natural human-computer interaction (2012, methodology): “We trained a multiclass SVM following the one-against-one method [Knerr et al. 1990] for fast training, while obtaining comparable accuracy to the 5:12 Y. Song et al. one-against-all method [Hsu and Lin 2002].”
Automatic Age Estimation Based on Facial Aging Patterns (2007, methodology): “SVM follows the 1-against-1 method [8] for multiclass classification and uses the RBF kernel with the bias 1.”
A Kernel-Based Feature Selection Method for SVM With RBF Kernel for Hyperspectral Image Classification (2014, methodology): “This study compared the classification performance based on the feature subsets by applying the proposed method KFS, BAHSIC [37], FS [19], FSFS [58], [59], and RFE [45], [60] for the SVM with a one-against-all (OAA) multiclass strategy [26], [27].”
A Unified View on Multi-class Support Vector Classification (2016, background): “This is in accordance with the ﬁndings of Hsu and Lin (2002) and Rifkin and Klautau (2004).”
Sleep Stage Classification Using EEG Signal Analysis: A Comprehensive Survey and New Investigation (2016, background): “For a detailed review and mathematical foundation of the six aforementioned classiﬁers, [130–133] can be reviewed.”
Efficient multi-label ranking for multi-class learning: Application to object recognition (2009, background): “The division usually follows one-vsall (OvA), one-vs-one or the general error correction code framework [6, 13, 11].”
Support vector machines for classification in remote sensing (2005, background): “The results of a recent analysis of multi-class strategies are provided by Hsu and Lin (2002).”
A tutorial review: Metabolomics and partial least squares-discriminant analysis--a marriage of convenience or a shotgun wedding. (2015, background): “The classification of the test set is determined by projecting each of the new points into this space and identifying to which side of the support vectors the new sample falls [32,34,111].”
An overview of ensemble methods for binary classifiers in multi-class problems: Experimental study on one-vs-one and one-vs-all schemes (2011, methodology): “Furthermore, binary decompositions have been widely used to develop multi-class SVM showing better performance than other multi-class SVM approaches [42].”