A cluster-preserving feature selection framework cut intrusion detection features from 78 to 27 on the CICIDS2017 benchmark while Random Forest and XGBoost maintained roughly 99.9 percent accuracy.
A new multi-granularity fuzzy rough feature selection method fuses algebraic and informational views, improving classification accuracy across 20 benchmark datasets.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results