Boosting for Learning Multiple Classes with Imbalanced Class Distribution

By Yanmin Sun, Mohamed S. Kamel, Yang Wang. Industrial Conference on Data Mining, 2006

322 citations recorded by the source corpus.

Subfields: Accounting, Artificial Intelligence, Electrical and Electronic Engineering

View Paper: https://www.semanticscholar.org/paper/c92017adc02a6a4887a9c65fe16b7f6d6ca89e61

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

Learning from Imbalanced Data (2009, background): “Although this description would seem to imply that all between-class imbalances are innately binary (or two-class), we note that there are multiclass data in which imbalances exist between the various classes [7], [8], [9], [10], [11], [12].”
A Survey of Predictive Modeling on Imbalanced Domains (2016, methodology): “Sun et al. [2006] presented the MA v G metric, a generalization of the G-Mean for more than two classes (cf. Equation (27)).”
Comparing Random Forest with Logistic Regression for Predicting Class-Imbalanced Civil War Onset Data (2016, methodology): “Existing techniques include sampling methods, like downsampling, which we utilize in this article (Ling and Li 1998; Chawla et al. 2002; Cieslak and Chawla 2008; Ko¨knar-Tezel and Latecki 2011), and ensemble-based methods (Chawla et al. 2003; Sun et al. 2006), among others.”
On multi-class classification through the minimization of the confusion matrix norm (2013, methodology): “Usual performance measures of multi-class classification include MAUC (an extension of the AUC to multi-class problems, (Hand and Till, 2001)), and G-mean (the geometric mean of recall values over all classes (Sun et al., 2006)).”
Multiclass Imbalance Problems: Analysis and Potential Solutions (2012, methodology): “To assess the overall performance across all classes, the generalized versions of AUC and G-mean, i.e., MAUC [26] and extended G-mean [19], are adopted in the following discussions.”
Financial credit risk assessment: a recent review (2015, background): “An advanced solution is cost-sensitive boosting that introduces cost items into boosting ensemble, such as AdaC1, AdaC2, and AdaC3 (Sun et al. 2006, 2007).”
A Graphbased Framework for Multi-Task Multi-View Learning (2011, methodology): “Existing methods can be roughly categorized into 3 groups [5], i.e. sampling based methods [21][19][6], adapting learning algorithms by modifying objective functions or changing decision thresholds [28][16], and ensemble based methods [27][7].”
Boosting methods for multi-class imbalanced data classification: an experimental review (2020, methodology): “Due to the independence of the imbalanced distribution, Geometric-mean ( G-mean) is another suitable single metric for imbalanced data problems [19] that is defined as Eq.”
Consistent Multiclass Algorithms for Complex Performance Measures (2015, methodology): “The G-mean measure is used to evaluate both binary and multiclass classifiers in settings with class imbalance (Sun et al., 2006; Wang & Yao, 2012), and is given by ψ GM (C) = n i=1 C ii n j=1 C ij 1/n .”
Combating the Small Sample Class Imbalance Problem Using Feature Selection (2010, methodology): “Sun et al. [26] used a genetic algorithm to identify the best cost matrix for an imbalanced problem and then applied a standard boosting algorithm to classify the data set.”
Boosting support vector machines for imbalanced data sets (2008, background): “However, it is important to keep in mind that the class imbalance problem is pervasive in other areas as well such as multi-class classification [37] and association rule mining [33].”
To Combat Multi-Class Imbalanced Problems by Means of Over-Sampling Techniques (2016, unknown): “[47] is provided in equation (8).”