An Agile Approach to Identify Single and Hybrid Normalization for Enhancing Machine Learning-Based Network Intrusion Detection

Detecting intrusion in network traffic has remained a problematic task for years. Progress the field of machine learning is paving way enhancing detection systems. Due to this progress become an integral part security. Intrusion achieved high accuracy with help supervised methods. A key factor performance classifiers how data augmented training classification model. Data real-world networks or publicly available datasets are not always normally (Gaussian) distributed. Instead, distributions variables more likely be skewed. To achieve rate, normalization transformation plays important role learning-based Several methods normalize attributes before However, opting most suitable technique still questionable task. In paper, statistical method proposed that can identify dataset. The identified by approach gives highest system. highlight efficiency method, five different were used two feature selection belong both Internet things and traditional environments. also able hybrid normalizations even improved results.