An Improved Perceptron Tree Learning Model Based Intrusion Detection Approach

摘要: This paper dedicates to develop an improved perceptron tree (PT) learning model based intrusion detection approach. The binary structure of a PT enables the divide problem into sub-problems and solve them in decreased complexity different levels. expert neural networks (ENNs) embedded internal nodes can be simplified by limiting number inputs hidden neurons. potential advantage is that trained actually “gray box” since each ENN interpreted explicit rules easily. However, whole likely high complex, i.e., probably composed large nodes. In this case, disjunctive description learned extracted from such too complex understand. generalization ability approach may depressed as well. view situation, needs fine pruned. experimental results demonstrate proposed achieve competitive accuracy well refined structure.