Intrusion Detection system using machine learning (Random Forest)
The dataset used is Canadian Institute for cyber security intrusion detection system (CICIDS-2017) which includes 76 features with 5 types of attack and those are ('PortScan', 'BENIGN', 'DDoS', 'Web Attack XSS', 'Bot'). Random forest is applied for the detection of the attack and different feature selection algorithm is used to clean the data having correlated features, constant and quasi constant features, duplicate features and misiing values as well as -inf and inf values. Further different performance metric is used for prediction and analysis.