Using Incremental Mining to Generate Fuzzy Rules for Real-Time Network Intrusion Detection Systems

摘要

In the paper, we propose a fast algorithm to generate fuzzy association rules by incremental mining approach, for which the transactions or data records are online instantly collected from live packets. That is, as one data record is collected online, the latest fuzzy rules can be obtained immediately. According to our simulation, in case of the number of features do not excess twenty, mining process can be completed from several milliseconds to seconds depending on the number of features.