The Foundations of Cost-Sensitive Learning

By Charles Peter Elkan. 2001

2,681 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 Optimization & Training

Citations in the corpus

Cascade R-CNN: High Quality Object Detection and Instance Segmentation (2019, background): “This is known in the cost-sensitive learning literature [11], [41], where the optimization of different points of the receiver operating characteristic (ROC) requires different loss functions.”
Cascade R-CNN: Delving Into High Quality Object Detection (2017, background): “This is known in the cost-sensitive learning literature [7, 26], where the optimization of different points of the receiver operating characteristic (ROC) requires different loss functions.”
Three naive Bayes approaches for discrimination-free classification (2010, background): “In Cost-Sensitive and Utility-Bases learning (Chan and Stolfo 1998; Elkan 2001; Margineantu and Dietterich 1999), it is assumed that not all types of prediction errors are equal and not all examples are as important.”
Detecting and Correcting for Label Shift with Black Box Predictors (2018, background): “When q ( y ) is known, label shift simpliﬁes to the problem of changing base rates (Bishop, 1995; Elkan, 2001).”
Domain Adaptation: Learning Bounds and Algorithms (2009, methodology): “This bears some similarity with the reweighting or importance weightingideas used in statistics and machine learning for sample bias correction techniques (Elkan, 2001; Cortes et al., 2008) and other purposes.”
The Devil is in the Tails: Fine-grained Classification in the Wild (2017, methodology): “In the ﬁrst regime, cost-sensitive learning [10] is employed to force the model to adjust its decision boundaries by incurring a non-uniform cost per misclassiﬁcation; see [19] for a review of the techniques.”
Class-Balanced Loss Based on Effective Number of Samples (2019, background): “[10] studied how to assign weights to adjust the decision boundary to match a given target in the case of binary clas-siﬁcation.”
A systematic study of the class imbalance problem in convolutional neural networks (2017, methodology): “This method assigns diﬀerent cost to misclassiﬁcation of examples from diﬀerent classes [44].”
Finding a short and accurate decision rule in disjunctive normal form by exhaustive search (2010, positive): “Cost-insensitive algorithms can be made cost-sensitive by resampling or wrapper techniques (Elkan 2001; Viaene and Dedene 2005), such as MetaCost (Domingos 1999).”
Learned lessons in credit card fraud detection from a practitioner perspective (2014, background): “The cost of a fraud is often assumed to be equal to the transaction amount (Elkan, 2001).”
A Survey of Predictive Modeling on Imbalanced Domains (2016, background): “Regarding further theoretical contributions, we must highlight that this equivalence was further explored by Elkan [2001].”
Learning with Rejection (2016, background): “Another somewhat similar approach is that of cost-sensitive learning where a class-dependent cost can be used [13].”
The cost of fairness in binary classification (2018, background): “For the costsensitive risk with parameter c, the Bayes-optimal classifier is f ∗(x) = nη(x) > co + α · nη(x) = co (Elkan, 2001), where η(x) .”
Learning from positive and unlabeled data: a survey (2018, methodology): “5c [24].”