INTRUSION DETECTION USING GRNN AND RANDOM FOREST

Along with the number of people building computer networks connected to the internet, then increasingly potential to cyber threats such as network intrusion (interference on the network). What includes intrusion in computer networks is the act of trying bypassing computer network system security mechanisms. One attempt to detect intrusion in the network is to differentiate network traffic activity. To distinguish normal network traffic activity with abnormal is difficult and tedious. Network analysts must examine all large and wide-ranging data to find the order anomalies (odd) on the network connection. GRNN and Random Forest can be used to group events on the network based on attributes. Each event on the network will be derived into a unique section by the decision tree. Order events on the network are mapped to the sequence of connected sections. By building rules based on part order generate intrusion alerts that can detect any attempt to do an intrusion. However, using the GRNN and Random forest over CICIDS2017 dataset the scheme is able to achieve 79.37% of accuracy over the cyber attacks.