Cost-sensitive boosting for classification of imbalanced data

By Yanmin Sun, Mohamed S. Kamel, Andrew K. C. Wong, Yang Wang. Pattern Recognition, 2007

1,536 citations recorded by the source corpus.

Subfields: Accounting, Artificial Intelligence, Electrical and Electronic Engineering

View Paper: https://www.semanticscholar.org/paper/b1e4660c82951ba19a861afc8be5cf56cc46aaa8

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

Analysing the classification of imbalanced data-sets with multiple classes: Binarization techniques and ad-hoc approaches (2013, background): “Cost-sensitive solutions: These incorporate approaches at data level, at algorithmic level, or at both levels jointly, considering higher misclassiﬁcation costs for the examples of the positive class with respect to the negative class, and therefore, trying to minimise higher cost errors…”
Training and assessing classification rules with imbalanced data (2012, background): “Some references are Liu et al. (2006); Thomas et al. (2006); Sun et al. (2007); Khoshgoftaar et al. (2007, 2011); Fernandez et al. (2012).”
Adaptive Machine Learning for Credit Card Fraud Detection (2015, methodology): “Example of these classifiers are cost-sensitive boosting [125, 126], SVM [127] and Neural Network [128].”
Comparing and experimenting machine learning techniques for code smell detection (2015, background): “Furthermore, the problem of the imbalance data has only recently gained wider attention in the machine learning community, e.g., (He and Garcia, 2009) and (Sun et al, 2007).”
A Scenario-Adaptive Driving Behavior Prediction Approach to Urban Autonomous Driving (2017, methodology): “Bearing in mind that most events belong to LK, evaluation measures for the imbalance data [38] were selected to evaluate the performance of Approaches A and B, i.”
Analysis of preprocessing vs. cost-sensitive learning for imbalanced classification. Open problems on intrinsic data characteristics (2012, methodology): “For this reason we also consider a hybrid approach in this work, ‘‘SMOTE + ENN’’, where the Wilson’s ENN rule (Wilson, 1972) is used after the SMOTE application to remove from the training set any example misclassified by its three nearest neighbors.”
A 30-m landsat-derived cropland extent product of Australia and China using random forest machine learning algorithm on Google Earth Engine cloud computing platform (2018, positive): “While training sample imbalance can a ﬀ ect the RF classi ﬁ cation output by over-ﬁ tting the majority class (Breiman, 2001; Chen et al., 2004), various methods such as down-sampling the majority class or duplicating the minority samples can provide immunity against over-ﬁ tting (Sun et al., 2007).”
CrisisLex: A Lexicon for Collecting and Filtering Microblogged Communications in Crises (2014, methodology): “…mean of the recall of the positive class and the recall of the negative class – often used to assess the classification performance on imbalanced data (Sun et al. 2007) After tokenizing, we discard tokens that are too short (2 characters or less), too long (16 characters or more, typically…”
A novel ensemble method for classifying imbalanced data (2015, background): “The classifiers built with these algorithms for such problems usually ignore the minority class as these classification algorithms tend to maximize the overall classification accuracy [10,18].”