Applying Support Vector Machines to Imbalanced Datasets

By Rehan Akbani, Stephen Kwek, Nathalie Japkowicz. European Conference on Machine Learning, 2004

1,285 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Information Systems

View Paper: https://www.semanticscholar.org/paper/45a9e2aa04e91bb511f36c365ef4daa274fe583c

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 Information Systems

Citations in the corpus

Imbalanced Deep Learning by Minority Class Incremental Rectification (2018, methodology): “Another common approach is cost-sensitive learning, which reformulates existing learning algorithms by weighting the minority classes more [8,9,10].”
Support vector machines in DSC‐based glioma imaging: Suggestions for optimal characterization (2010, methodology): “Machine learning algorithms, also SVMs, are prone to such imbalanced classes resulting in biased results, in most cases, the separating hyper plane is optimized toward the over-represented class (40,41).”
A review of ensemble learning and data augmentation models for class imbalanced problems: combination, implementation and evaluation (2023, background): “A disadvantage of SVM is that although it performs well on balanced datasets, its performance drops sharply on imbalanced datasets [224].”
A Survey of Predictive Modeling on Imbalanced Domains (2016, background): “Regarding support vector machines, several ways of integrating costs have been considered, such as assigning different penalties to false negatives and positives [Akbani et al. 2004] or including a weighted attribute strategy [Yuanhong et al. 2009], among others [Weiguo et al. 2012].”
Bayesian mixed-effects inference on classification performance in hierarchical data sets (2012, methodology): “Finally, both above methods assess performance in terms of accuracy, which may lead to inflated estimates for imbalanced data sets and thus to false conclusions about the significance with which the algorithm has performed better than chance (Chawla et al., 2002; Japkowicz and Stephen, 2002; Akbani ”
Cost-sensitive boosting for classification of imbalanced data (2007, background): “These classi-ﬁcation systems involve decision trees [1,8–10], support vector machines [9,11–13], neural networks [9], Bayesian network [6], nearest neighbor [8,14] and the newly reported associative classiﬁcation approaches [15,16].”
Disentangling Label Distribution for Long-tailed Visual Recognition (2020, methodology): “Most long-tailed visual recognition methods can be divided into two strategies: modifying the data sampler to balance the class frequency during optimization [7, 21, 25, 8, 51], and modifying the class-wise weights of the classification loss to increase the importance of tail classes in terms of emp”
Borderline over-sampling for imbalanced data classification (2009, background): “In fact, however, SVMs are still affected by the data imbalance to some extent [1].”
Weighted extreme learning machine for imbalance learning (2013, background): “Typically, to deal with imbalanced data one can attempt to establish balanced data distribution through various sampling methods or algorithmic approaches [7,8].”