arxivcs.CRcs.LG2026-07-14
BARS: Benign-Anchored Ranking and Selection for False Alarm Reduction in Network Intrusion Detection
Abu Fuad Ahmad, Istiaque Ahmed
False alarms remain a major barrier to deploying network intrusion detection systems (NIDS). In high-volume environments, even a sub-1% false positive rate can generate tens of thousands of daily alerts. Filter-based feature selection is attractive because it operates upstream of…