Efficient Pairwise Classification

Park, Sang-Hyeun; Fürnkranz, Johannes

doi:10.1007/978-3-540-74958-5_65

Cited by 70 publications

(43 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Finally, we are also working on expanding these results to other learning problems where binary decomposition techniques are applied. To that end, we have obtained very positive results for the calibrated ranking approach to multilabel classification (Mencía et al 2010), and have promising preliminary results for ranking problems (Park and Fürnkranz 2007b). a subset of non-incident classifiers of c i .…”

Section: Discussionmentioning

confidence: 89%

See 1 more Smart Citation

Efficient prediction algorithms for binary decomposition techniques

Park

Fürnkranz

2011

Data Min Knowl Disc

Self Cite

View full text Add to dashboard Cite

Binary decomposition methods transform multiclass learning problems into a series of two-class learning problems that can be solved with simpler learning algorithms. As the number of such binary learning problems often grows super-linearly with the number of classes, we need efficient methods for computing the predictions. In this paper, we discuss an efficient algorithm that queries only a dynamically determined subset of the trained classifiers, but still predicts the same classes that would have been predicted if all classifiers had been queried. The algorithm is first derived for the simple case of pairwise classification, and then generalized to arbitrary pairwise decompositions of the learning problem in the form of ternary error-correcting output codes under a variety of different code designs and decoding strategies.

show abstract

Section: Discussionmentioning

confidence: 89%

“…The same experiments were already performed without parameter tuning and can be found in (Park and Fürnkranz 2007a). Table 1 shows the results.…”

Section: Uci Datasetsmentioning

confidence: 99%

Efficient prediction algorithms for binary decomposition techniques

Park

Fürnkranz

2011

Data Min Knowl Disc

Self Cite

View full text Add to dashboard Cite

show abstract

“…We still have to store a quadratic number of classifiers, and, in principle, all of them have to be queried at classification time. Park and Fürnkranz (2007) have recently proposed an efficient algorithm that allows to compute the top-ranked class for prediction in essentially linear time. We are currently working on an adaptation of this algorithm for multilabel classification, which will compute the ranking from top to bottom, and stop as soon as the calibration label has been found.…”

Section: Computational Complexitymentioning

confidence: 99%

Multilabel classification via calibrated label ranking

et al. 2008

Self Cite

View full text Add to dashboard Cite

Label ranking studies the problem of learning a mapping from instances to rankings over a predefined set of labels. Hitherto existing approaches to label ranking implicitly operate on an underlying (utility) scale which is not calibrated in the sense that it lacks a natural zero point. We propose a suitable extension of label ranking that incorporates the calibrated scenario and substantially extends the expressive power of these approaches. In particular, our extension suggests a conceptually novel technique for extending the common learning by pairwise comparison approach to the multilabel scenario, a setting previously not being amenable to the pairwise decomposition technique. The key idea of the approach is to introduce an artificial calibration label that, in each example, separates the relevant from the irrelevant labels. We show that this technique can be viewed as a combination of pairwise preference learning and the conventional relevance classification technique, where a separate classifier is trained to predict whether a label is relevant or not. Empirical results in the area of text categorization, image classification and gene analysis underscore the merits of the calibrated model in comparison to state-of-the-art multilabel learning methods.

show abstract

“…Generalization of QWEIGHTED (Park and Fürnkranz, 2007), a simple efficient voting-based pairwise prediction algorithm, to ternary ECOC decoding guaranteed to produce the same result as with original decoding QWEIGHTED basic idea: not all pairwise comparisons are needed for determining only the winner class imagine the voting process as a tournament of two-player games consider dominant player who can not be caught up algorithm in one sentence: Evaluate always the classifier of the two best classes (w.r.t voting) until the best class has no evaluations left.…”

Section: Efficient Decoding Of Ecocmentioning

confidence: 99%