A Survey of Predictive Modeling on Imbalanced Domains

By Paula Branco, Luís Torgo, Rita P. Ribeiro. ACM Computing Surveys, 2016

1,064 citations recorded by the source corpus.

Many real-world data-mining applications involve obtaining predictive models using datasets with strongly imbalanced distributions of the target variable. Frequently, the least-common values of this target variable are associated with events that are highly relevant for end users (e.g., fraud detection, unusual returns on stock markets, anticipation of catastrophes, etc.). Moreover, the events may have different costs and benefits, which, when associated with the rarity of some of them on the available training data, creates serious problems to predictive modeling techniques. This article presents a survey of existing techniques for handling these important applications of predictive analytics. Although most of the existing work addresses classification tasks (nominal target variables), we also describe methods designed to handle similar problems within regression tasks (numeric target variables). In this survey, we discuss the main challenges raised by imbalanced domains, propose a definition of the problem, describe the main approaches to these tasks, propose a taxonomy of the methods, summarize the conclusions of existing comparative studies as well as some theoretical analyses of some methods, and refer to some related problems within predictive modeling.

Subfields: Artificial Intelligence

View Paper: https://www.semanticscholar.org/paper/736018160f8e6f6bd9b2a9101cd13ee67e573097

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

On Cross-Dataset Generalization in Automatic Detection of Online Abuse (2020, background): “Task formulation bias: In commercial applications, the definitions of offensive language heavily rely on community norms and context and, therefore, are imprecise, application-dependent, and constantly evolving (Chandrasekharan et al., 2018).”
Deep Neural Networks Improve Radiologists’ Performance in Breast Cancer Screening (2019, methodology): “First, while we trained the cancer classification model on data from all screening exams, within each training epoch, the model was shown all exams with biopsies in the training set (4,844 exams) but only a random subset of an equal number of exams without biopsies (also 4,844 exams) (4).”
Unequal-Training for Deep Face Recognition With Long-Tailed Noisy Data (2019, background): “…in face recognition is reminiscent of the conventional class imbalance problem that has been comprehensively studied in classical machine learning [2,10], but signiﬁcantly differs from conventional class imbalance problem in two aspects: ﬁrst, the long-tailed data in face recognition is…”
Ensemble Learning for Disease Prediction: A Review (2023, methodology): “We followed a thorough literature search approach to discover all relevant publications in this study.”
Online Continual Learning from Imbalanced Data (2020, background): “A well-known and effective way to deal with such imbal ances is oversampling the minority classes (Branco et al., 2016).”
Multiple instance learning: A survey of problem characteristics and applications (2016, methodology): “There exist many methods to deal with imbalanced data [155].”
SMOGN: a Pre-processing Approach for Imbalanced Regression (2017, methodology): “The introduction of Gaussian Noise (Branco et al., 2016a) is an adaptation to regression of the method proposed by Lee (1999, 2000) for classification tasks. This method also combines random under-sampling of the normal cases with the generation of synthetic rare examples. However, the new cases are”