Evaluating boosting algorithms to classify rare classes: comparison and improvements

By Mahesh V. Joshi, Vipin Kumar, Ramesh C. Agarwal. Proceedings 2001 IEEE International Conference on Data Mining, 2001

311 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/3f43487929447fe8ec76885dfc25d9f5439833b0

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

Citations in the corpus

Learning When Data Sets are Imbalanced and When Costs are Unequal and Unknown (2003, background): “There have also been proposals to boost the examples of the minority class (Joshi et al., 2001).”
To Combat Multi-Class Imbalanced Problems by Means of Over-Sampling Techniques (2016, background): “Bagging [5], Boosting [22], SMOTEBoost [9], and RareBoost [33] are popular in literature.”
Adaptive Machine Learning for Credit Card Fraud Detection (2015, methodology): “RareBoost [123] modifies the boosting algorithm to increase accuracy on the rare class by emphasizing the difference”
Data Mining for Imbalanced Datasets: An Overview (2005, background): “On the other side, Rare-Boost (Joshi et al., 2001) updates the weights of the examples differently for all four entries shown in Figure 40.”
Outlier Analysis (2013, background): “An evaluation of boosting algorithms for rare class detection is provided in [246].”
An insight into classification with imbalanced data: Empirical results and current trends on using data intrinsic characteristics (2013, methodology): “Following this idea, many approaches have been developed by modifying the standard boosting weight-update mechanism in order to improve the performance on the minority class and the small disjuncts [30,44,61,69,74,112,117,122].”
Learning from imbalanced data sets with boosting and data generation: the DataBoost-IM approach (2004, background): “It follows that the F-measure is high when both the recall and precision are high [6].”
A systematic study of the class imbalance problem in convolutional neural networks (2017, background): “The measure that takes into account precision and recall is F1-measure [96].”
Comparing Boosting and Bagging Techniques With Noisy and Imbalanced Data (2011, methodology): “In addition to data sampling, various algorithms derived from boosting methods have been proposed for handling class imbalance (or cost-sensitive problems), including AdaCost [20], CSB0, CSB1, and CSB2 [21], AdaC1, AdaC2, and AdaC3 [22], RareBoost [23], SMOTEBoost [10], and RUSBoost [11].”
Borderline-SMOTE: A New Over-Sampling Method in Imbalanced Data Sets Learning (2005, methodology): “The new algorithm can achieve better prediction for the minority class [19].”
Mining with rarity: a unifying framework (2004, background): “Because AdaCost, unlike RareBoost, does not stratify these measures separately, it is believed that AdaCost may sometimes over-emphasize recall, thus leading to poorer precision.”
RUSBoost: A Hybrid Approach to Alleviating Class Imbalance (2010, background): “Several variations have been proposed to make AdaBoost cost sensitive [15]–[17] or to improve its performance on imbalanced data [18]–[20].”
Learning from Imbalanced Data (2009, background): “Many representative works on the ineffectiveness of accuracy in the imbalanced learning scenario exist in the community [14], [20], [47], [51], [58], [116], [117], [118].”
Cost-sensitive boosting for classification of imbalanced data (2007, methodology): “However, for classification of imbalanced data, accuracy is no longer a proper measure since the rare class has very little impact on the accuracy as compared to that of the prevalent class [47, 99].”
Data Mining for Network Intrusion Detection (2002, methodology): “When our proposed association based classification algorithm is applied on KDDCup data set, experimental results indicate that the significant increase in prediction performance may be achieved by considering not only the itemsets with high precision but also the itemsets with high recall and F-valu”