Classification of Imbalanced Data: a Review

By Yanmin Sun, Andrew K. C. Wong, Mohamed S. Kamel. International journal of pattern recognition and artificial intelligence, 2009

1,588 citations recorded by the source corpus.

Subfields: Accounting, Artificial Intelligence, Electrical and Electronic Engineering

View Paper: https://www.semanticscholar.org/paper/e2c21999756ef8e21cdd931b780ed1a80ca09a2d

How researchers see this paper

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

In Accounting

In Artificial Intelligence

In Electrical and Electronic Engineering

Citations in the corpus

On the use of MapReduce for imbalanced big data using Random Forest (2014, background): “The obtaining of the cost matrix is a key issue when using cost-sensitive learning and should be provided by domain experts, however, when it is not available an estimation can be used [49,50].”
Training and assessing classification rules with imbalanced data (2012, background): “Providing a complete review of the inherent literature is beyond the scope of this paper (for two excellent reviews, see, e.g. He and Garcia 2009; Sun et al. 2009).”
Image Data Augmentation for Deep Learning: A Survey (2022, background): “5 Discussion for Future Directions et al. , 2009].”
Learning from imbalanced data: open challenges and future directions (2016, background): “Binary classification problems can be considered as the most developed branch of learning from imbalanced data [51].”
An overview of ensemble methods for binary classifiers in multi-class problems: Experimental study on one-vs-one and one-vs-all schemes (2011, background): “…usually imbalanced training data sets are produced when instances from the single class are compared with all other instances in the data set, it is well-known in the field of Machine Learning that imbalanced data sets can cause some undesirable effects in the derived classifiers [15] [40] [70].”
A review of microarray datasets and applied feature selection methods (2014, background): “This occurs when a dataset is dominated by a major class or classes which have significantly more instances than the other rare/minority classes in the data [56,116,76].”
A Systematic Review on Imbalanced Data Challenges in Machine Learning (2019, background): “(3) The next challenge in skewed data distribution is uncertain method of selecting samples (Sun et al. 2009) (4) Other challenges include different distributions in various subclasses contained in a single class as it results in incrementing the learning complexity of the dataset.”
Analysing the classification of imbalanced data-sets with multiple classes: Binarization techniques and ad-hoc approaches (2013, background): “The main dif-ﬁculty in approaching this problem is that standard learning algorithms consider a balanced training set which induces a bias towards the majority classes [53].”
A Survey of Predictive Modeling on Imbalanced Domains (2016, background): “Problems such as small disjuncts, class overlap, and small sample size usually coexist with imbalanced classiﬁcation domains and are also identiﬁed as possible causes of classiﬁers performance degradation [Weiss 2004; He and Garcia 2009; Sun et al. 2009; Stefanowski 2016].”
Cost-sensitive decision tree ensembles for effective imbalanced classification (2014, positive): “Class imbalance not only makes the learning task more complex [38], it is usually accompanied also by other difficulties such as: • Small sample size: In many cases the number of minority class samples is insufficient to properly train a classifier, hence resulting in poor generalisation and possibl”
EUSBoost: Enhancing ensembles for highly imbalanced data-sets by evolutionary undersampling (2013, background): “Even though the uneven class distribution need not be a problem itself, this situation usually produces overlapping or small disjuncts, added to the frequently small sample size of the minority class [50,25].”