Cost-sensitive learning by cost-proportionate example weighting

By Bianca Zadrozny, J. Langford, Naoki Abe. 2003

758 citations recorded by the source corpus.

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

Citations in the corpus

Preventing Fairness Gerrymandering: Auditing and Learning for Subgroup Fairness (2017, background): “Cost-sensitive classification is polynomially equivalent to agnostic learning (Zadrozny et al., 2003).”
The offset tree for learning with partial labels (2008, methodology): “The importance weighted regret is bounded by the unweighted regret, times the expected importance (see [22]), which in turn is bounded by k.”
An Algorithmic Perspective on Imitation Learning (2018, methodology): “Model-Free Behavioral Cloning Methods in Action-State space61 SEARN learns a multiclass cost-sensitive classifier, e.g. [Zadrozny et al., 2003], for each state in the dataset through an iterative process.”
Modelling input texts: from Tree Kernels to Deep Learning (2015, methodology): “Our approach to accomplish cost-sensitive classification shares the idea of reductions put forward in [190] together with the benefit of the conventional approach in SVMs [168] to incorporate importance weights directly into the”
Relating reinforcement learning performance to classification performance (2005, result): “Experimental results for some simple reductions (Zadrozny et al., 2003; Beygelzimer et al., 2004; Langford & Zadrozny, 2005) suggest that this is true, but more work needs to be done.”
Data mining with imbalanced class distributions: concepts and methods (2009, methodology): “Non näıve approaches use elaborated sampling techniques [24] or a combination of multiple classifiers induced with different class proportions [25], although the general idea is to induce a cost sensitive classifier by directly or indirectly (using weights) changing the class proportion.”
Editorial: special issue on learning from imbalanced data sets (2004, methodology): “He also discussed a cost-sensitive ensemble learning method called Costing [59], which achieved significant improvements over random resampling methods.”
Sample Selection Bias Correction Theory (2008, methodology): “This weight is used to emphasize or de-emphasize the cost of an error on z i as in the so-called importance weighting or cost-sensitive learning (Elkan, 2001; Zadrozny et al., 2003).”
Learning using privileged information: SV M+ and weighted SVM (2013, background): “We only give a few pointers to papers on cost-sensitive learning (Margineantu, 2002; Zadrozny et al., 2003), sample bias correction (Heckman, 1979; Cortes et al., 2010), domain adaptation (Shimodaira, 2000; Sugiyama & Müller, 2005), online learning (Dredze et al., 2008), and active learning…”
An insight into classification with imbalanced data: Empirical results and current trends on using data intrinsic characteristics (2013, methodology): “The most popular technique lies in resampling the original class distribution of the training dataset according to the cost decision matrix by means of undersampling/oversampling [148] or assigning instance weights [123].”
Learning Deep Representation for Imbalanced Classification (2016, background): “To counter the negative effects, one often chooses from a few available options, which have been extensively studied in the past [7, 9, 11, 17, 18, 30, 40, 41, 46, 48].”
Outlier Analysis (2013, methodology): “Some methods have also been proposed in the literature [496] in order to incorporate the weights explicitly into the learning process.”
Imbalanced Deep Learning by Minority Class Incremental Rectification (2018, methodology): “(2) Algorithmlevel: Modifying existing algorithms to give more emphasis on the minority classes [8,9,10,43,44,45,46,47,48,49,50].”
Get another label? improving data quality and data mining using multiple, noisy labelers (2008, background): “Moreover, any importance-weighted classiﬁcation problem can be reduced to a uniform-weighted classiﬁcation problem [35], often performing better than hand-crafted weighted-classiﬁcation algorithms.”