Learning from Imbalanced Data Sets: A Comparison of Various Strategies *

By Nathalie Japkowicz. 2000

553 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/4c6e99daaac3e78dc12864102f8072a5ef8278e1

How researchers see this paper

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

Citations in the corpus

A study on several Machine-learning methods for classification of Malignant and benign clustered microcalcifications (2005, background): “Aimed to improving the classification performance, several strategies exist for dealing with unbalanced data sets for classifier training, including resizing the training data sets, adjusting misclassification costs, and recognition-based learning (learning from the minority class) [38].”
Customer churn prediction in telecommunications (2012, background): “In addition, because there is usually a very large number of nonchurners and a very small numbers of churners, the class imbalance problem (Japkowicz, 2000) occurs in this application.”
Learning When Data Sets are Imbalanced and When Costs are Unequal and Unknown (2003, background): “There have been several proposals for coping with skewed data sets (Japkowicz, 2000).”
An experimental comparison of classification algorithms for imbalanced credit scoring data sets (2012, methodology): “In Japkowicz (2000), over-sampling and downsizing were compared to the author’s own method of ‘‘learning by recognition’’ in order to determine the most effective technique.”
Semi-automated screening of biomedical citations for systematic reviews (2010, background): “This approach has been shown to work well with respect to increasing classifier sensitivity to the minority class [30].”
Experimental perspectives on learning from imbalanced data (2007, unknown): “…evaluated the use of sampling when learning from imbalanced data (e.g., (Drummond & Holte, 2003), (Kubat & Matwin, 1997), (Malo of, 2003), (Japkowicz, 2000), (Weiss & Provost, 2003), (Chawla et al., 2002), (Han et al., Appearing in Proceedings of the 24 th International…”
A Survey of Predictive Modeling on Imbalanced Domains (2016, background): “…made in this context include one-class SVMs (e.g., Sch¨olkopf et al. [2001], Manevitz and Yousef [2002], Raskutti and Kowalczyk [2004], Zhuang and Dai [2006a, 2006b], and Lee and Cho [2006]) and the use of an autoencoder (or autoassociator) (e.g., Japkowicz et al. [1995] and Japkowicz [2000]).”
Data Mining for Imbalanced Datasets: An Overview (2005, methodology): “Over and under-sampling methodologies have received significant attention to counter the effect of imbalanced data sets (Solberg and Solberg, 1996; Japkowicz , 2000a; Chawla et al., 2002; Weiss and Provost, 2003; Kubat and Matwin, 1997; Jo and Japkowicz, 2004; Batista et al., 2004; Phua and…”
RUSBoost: A Hybrid Approach to Alleviating Class Imbalance (2010, background): “Japkowicz [9] presents another study addressing the issue of class imbalance, including an investigation of the types of imbalance that most negatively impact classification performance, and a small case study comparing several techniques for alleviating the problem.”
Learning to Recombine and Resample Data for Compositional Generalization (2020, background): “In classiﬁcation problems with imbalanced classes, a common strategy for improving accuracy on the rare class is to resample so that the rare class is better represented in training data (Japkowicz et al., 2000).”
Neighbor-weighted K-nearest neighbor for unbalanced text corpus (2005, background): “However, as indicated in (Japkowicz, 2000) unbalanced data sets often appear in many practical applications.”
Leveraging molecular structure and bioactivity with chemical language models for de novo drug design (2023, methodology): “To mitigate the class data imbalance, we applied over-sampling to the classes with fewer data (i.e., the “ inactive ” and “ highly active ” classes) 44 .”
Learning from imbalanced data sets with boosting and data generation: the DataBoost-IM approach (2004, background): “The class imbalance problem corresponds to domains for which one class is represented by a large number of examples while the other is represented by only a few [1].”
Imbalanced Deep Learning by Minority Class Incremental Rectification (2018, background): “Down-sampling, on the other hand, throws away valuable information [5,42].”
A methodological approach to the classification of dermoscopy images (2007, background): “Several studies have demonstrated that the accuracy degradation on unbalanced data sets is more severe when the classes overlap significantly [ 47 ,48]; this is the case in skin lesion classification.”
Boosting methods for multi-class imbalanced data classification: an experimental review (2020, background): “A data set is imbalanced when the samples of one class consist of more instances than the rest of the classes in two-class and multi-class data sets [1].”
On the Class Imbalance Problem (2008, methodology): “A direct method to solve the imbalance problem is artificially balancing the class distributions, and its effectiveness has been empirically analyzed in [11].”
The effect of class distribution on classifier learning: an empirical study (2001, background): “Kubat and Matwin (1997) developed an under-sampling strategy that intelligently removes majority examples by removing only those majority examples that are “redundant” or that “border” the minority examples—figuring they may be the result of noise.”