Balancing the accuracy rates of the majority and minority classes is challenging in imbalanced classification. Furthermore, data characteristics have a significant impact on the performance of imbalanced classifiers, which are generally neglected by existing evaluation methods. The objective of this study is to introduce a new criterion to comprehensively evaluate imbalanced classifiers. Specifically, we introduce an efficiency curve that is established using data envelopment analysis without explicit inputs (DEA-WEI), to determine the trade-off between the benefits of improved minority class accuracy and the cost of reduced majority class accuracy. In sequence, we analyze the impact of the imbalanced ratio and typical imbalanced data characteristics on the efficiency of the classifiers. Empirical analyses using 68 imbalanced data reveal that traditional classifiers such as C4.5 and the k-nearest neighbor are more effective on disjunct data, whereas ensemble and undersampling techniques are more effective for overlapping and noisy data. The efficiency of cost-sensitive classifiers decreases dramatically when the imbalanced ratio increases. Finally, we investigate the reasons for the different efficiencies of classifiers on imbalanced data and recommend steps to select appropriate classifiers for imbalanced data based on data characteristics.

Data Science and Machine Learning have become fundamental assets for companies and research institutions alike. As one of its fields, supervised classification allows for class prediction of new samples, learning from given training data. However, some properties can cause datasets to be problematic to classify. In order to evaluate a dataset a priori, data complexity metrics have been used extensively. They provide information regarding different intrinsic characteristics of the data, which serve to evaluate classifier compatibility and a course of action that improves performance. However, most complexity metrics focus on just one characteristic of the data, which can be insufficient to properly evaluate the dataset towards the classifiers' performance. In fact, class overlap, a very detrimental feature for the classification process (especially when imbalance among class labels is also present) is hard to assess. This research work focuses on revisiting complexity metrics based on data morphology. In accordance to their nature, the premise is that they provide both good estimates for class overlap, and great correlations with the classification performance. For that purpose, a novel family of metrics have been developed. Being based on ball coverage by classes, they are named after Overlap Number of Balls. Finally, some prospects for the adaptation of the former family of metrics to singular (more complex) problems are discussed.