SVMs Modeling for Highly Imbalanced Classification

By Yuchun Tang, Yanqing Zhang, Nitesh V. Chawla, Sven Krasser. 2009

968 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Electrical and Electronic Engineering

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

In Electrical and Electronic Engineering

Citations in the corpus

Sampling Bias and Class Imbalance in Maximum-likelihood Logistic Regression (2011, background): “Since the 1980’s and recently across a variety of disciplines, there have been several attempts by researchers to answer the question: How does class imbalance affect the predictive capability of asymptotic classification algorithms such as MLLR?(Burez and Van den Poel, 2008; Cosslett, 1981a; Garcia”
Cost-sensitive learning methods for imbalanced data (2010, background): “Other researchers concentrate on changing the classifier internally, for example SVM, to deal with class imbalance such as [11], [12], and [13]; [14] uses ensemble learning to deal with class imbalance while [15] combines undersampling with ensemble methods; [16] focuses on incorporating different r”
Facing Imbalanced Data Recommendations for the Use of Performance Metrics (2013, background): “This approach targets the problem of skew by applying different cost matrices that describe the costs for misclassifying any particular data point [26], [8].”
An Attention-Based Deep Learning Approach for Sleep Stage Classification With Single-Channel EEG (2021, background): “Both MF1 and MGm are common metrics to evaluate the performance of the models on imbalanced datasets [42].”
Learning Deep Representation for Imbalanced Classification (2016, background): “For example, in [40] the classic SVM is made cost-sensitive to improve classification on highly skewed datasets.”
Outlier Analysis (2013, background): “Numerous cost-sensitive variations of different classifiers have been proposed along the lines of weighting, and include the Bayes classifier [496], nearest neighbor classifier [506], decision trees [450, 463], rule-based classifiers [245, 247] and SVM classifiers [443, 470].”
SMOTE-IPF: Addressing the noisy and borderline examples problem in imbalanced classification by a re-sampling method with filtering (2015, methodology): “These are k-NN [39], a Support Vector Machine (SVM) [13] [51], Repeated Incremental Pruning to Produce Error Reduction (RIPPER) [12] and PART [16].”
Data Science of the People , for the People , by the People : A Viewpoint on an Emerging Dichotomy (2015, methodology): “When predictive models trained under expected error make gross mistakes, they are usually on people with rare characteristics; there are learning approaches to deal with rare classes [48], but there is little work on data points with rare features [58, 5].”
A Systematic Review on Imbalanced Data Challenges in Machine Learning (2019, methodology): “Thus, different heuristics were incorporated in SVM modelling, which includes oversampling, under-sampling, and cost-sensitive learning by Tang et al. (2009).”
The Devil is in Classification: A Simple Framework for Long-tail Object Detection and Instance Segmentation (2020, methodology): “Performance drop on head classes is also observed in imbalanced classification [17], [38].”
Loss Max-Pooling for Semantic Image Segmentation (2017, methodology): “Such methods were recently investigated [7, 35, 44, 45] for deep learning, some of them following ideas previously applied in shallow learning methods like random forests [27, 28] or support vector machines [38, 42].”
The Devil is in Classification: A Simple Framework for Long-tail Instance Segmentation (2020, methodology): “…classiﬁcation, we ﬁrst consider incorporating several common strategies developed for long-tail classiﬁcation into current instance segmentation frameworks, including loss re-weighting [19],[38], adaptive loss adjustment (focal loss [30], class-aware margin loss [7]), and data re-sampling [17,37].”