Imbalanced Data Machine Learning

In real world the datasets are very often imbalanced in the  sense that the representations from different classes are not approximately equal and at least one class is under-represented relative to others. Such datasets arise quite often from applications including (but not limited to) fraud/intrusion detection, medical diagnosis/monitoring, bioinformatics, text categorization etc. The problem of imbalanced data is often associated with asymmetric costs of misclassifying elements of different classes. Our research focuses on devising suitable variants of the support vector (maximum margin) classifiers and boosting algorithms for providing acceptable solutions to imbalanced data classification problems without using severe computational overheads.