International Journal of Modern Engineering Research (IJMER)
Intrusion detection systems are applied to detect network intrusions identified from different sources. Anomaly and signature based schemes are used for the intrusion detection process. Signature based intrusion detection schemes uses the pre-defined signature collection for the detection process. The anomaly-based model detects the intrusions by learning the network transaction patterns. Feature selection schemes are used to reduce the network transaction features. Performance, time, accuracy and reliability are improved by the feature selection schemes. Current intrusion detection systems use the TCP/IP header information for the intrusion detection process. Network layer and transport layer attacks can be easily detected using TCP/IP header information's.