Intrusion Detection based Sample Selection for imbalanced data distribution

A new approach for imbalanced data classification based on data gravitation

“…However, capturing enough malicious traffic samples, such as attacks and virus traffic, for training is difficult. Many real-world classification tasks, such as medical diagnosis [44], fraud detection [14], finance risk management [7], network intrusion detection [9], stream classification [26], and bioinformatics [67], have similar diagnosis characteristics. In these imbalanced tasks, the minority class is usually more important than the majority class [18,68].…”