Cost-Sensitive Learning and the Class Imbalance Problem

By Charles X. F. Ling, Victor S. Sheng. 2008

347 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/9c4a953ed2cfc999eef0901d43097f9d2933005c

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

Adaptive Machine Learning for Credit Card Fraud Detection (2015, methodology): “Similarly, thresholding [131] allows using cost-insensitive algorithms for cost minimization via different classification thresholds.”
Addressing Class Imbalance in Federated Learning (2020, background): “, meta learning (Ling and Sheng 2008; Wang et al. 2020), model tuning (Pouyanfar et al.”
Boosting methods for multi-class imbalanced data classification: an experimental review (2020, background): “The most common approach is the cost-sensitive learning in which the learner is forced to correctly classify the minority class samples by setting a high cost to the misclassified samples of minority class [32].”
Twitter Sentiment Analysis (2015, methodology): “In order to account for this unbalance we trained a cost-sensitive SVM model (Ling and Sheng, 2007).”
Data mining with imbalanced class distributions: concepts and methods (2009, background): “by different misclassification costs, cost sensitive learning might be applied to overcome the class imbalance problem [18].”
GraphSMOTE: Imbalanced Node Classification on Graphs with Graph Neural Networks (2021, background): “Data-level approaches seek to make the class distribution more balanced, using over-sampling or downsampling techniques[8, 26]; algorithm-level approaches typically introduce different mis-classification penalties or prior probabilities for different classes [11, 22, 44]; and hybrid approaches [9, 2”
SMOTE for Regression (2013, methodology): “The existing approaches for the classification scenario can be cast into 3 main groups [5, 6]: i) change the evaluation metrics to better capture the application bias; ii) change the learning systems to bias their optimization process to the goals of these domains; and iii) sampling approaches that ”
A review of ensemble learning and data augmentation models for class imbalanced problems: combination, implementation and evaluation (2023, background): “…to deal with CI datasets, such as under-sampling majority class [14, 15, 16, 17], oversampling minority class [18, 19], cost-sensitive learning [20], one-class classification [21, 22, 23], and data augmentation [24] that focuses on creating synthetic data to support minority classes [25, 5,…”