Abstract
Since canonical machine learning algorithms assume that the dataset has equal number of samples in each class, binary classification became a very challenging task to discriminate the minority class samples efficiently in imbalanced datasets. For this reason, researchers have been paid attention and have proposed many methods to deal with this problem, which can be broadly categorized into data level and algorithm level. Besides, multi-class imbalanced learning is much harder than binary one and is still an open problem. Boosting algorithms are a class of ensemble learning methods in machine learning that improves the performance of separate base learners by combining them into a composite whole. This paper’s aim is to review the most significant published boosting techniques on multi-class imbalanced datasets. A thorough empirical comparison is conducted to analyze the performance of binary and multi-class boosting algorithms on various multi-class imbalanced datasets. In addition, based on the obtained results for performance evaluation metrics and a recently proposed criteria for comparing metrics, the selected metrics are compared to determine a suitable performance metric for multi-class imbalanced datasets. The experimental studies show that the CatBoost and LogitBoost algorithms are superior to other boosting algorithms on multi-class imbalanced conventional and big datasets, respectively. Furthermore, the MMCC is a better evaluation metric than the MAUC and G-mean in multi-class imbalanced data domains.
Highlights
Imbalanced data set classification is a relatively new research line within the broader context of machine learning studies, which tries to learn from the skewed data distribution
For the sake of clarity, it should be noted that the library of all algorithms were installed using the pip Python installer, e.g., sudo pip install xgboost, except MEBoost, SMOTEBoost, and AdaCosts, which their implemented python source codes are freely available at GitHub3 repository
The results prove that both multi-class area under the curve (MAUC) and multi-class Matthews correlation coefficient (MMCC) are more discriminant than G-mean
Summary
Imbalanced data set classification is a relatively new research line within the broader context of machine learning studies, which tries to learn from the skewed data distribution. Most of the standard machine learning algorithms show poor performance in this kind of datasets, because they tend to favor the majority class samples, resulting in poor predictive accuracy over the minority class [2]. Tanha et al J Big Data (2020) 7:70 important instances. They assume equal misclassification cost for all samples for minimizing the overall error rate. Learning from skew datasets becomes very important when many real-world classification problems are usually imbalanced, e.g. fault prediction [3], fraud detection [4], medical diagnosis [5], text classification [6], oil-spill detection in satellite images [7] and cultural modeling [8]. In software fault prediction, if the defective module is regarded as the positive class and non-defective module as negative, missing a defect (false negative) is much expensive than the false-positive error in testing phase of software development process [9]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
