Learning active instances on the border in the case of an imbalanced data classification task.
The implementation is based on the border learning: active learning algorithm for imbalance classification with early stopping by using small pools in active.py module.
An extension and transfer of the technique to the area of highly interpretable and robust prototype-based models LVQ's has been exemplified with matrix-lvq in the active1.py script.
Ertekin, Seyda, et al. "Learning on the border: active learning in imbalanced data classification." Proceedings of the sixteenth ACM conference on Conference on information and knowledge management. 2007.