A weighted decision tree-based fast intrusion detection model

From people's cognition process, we propose a model to solve network invading problem in real time. It abstracts network traffic features in the order of network data which arrives in reality, computes and appraises network features then makes a decision whether to report. In order to improve the reality of intrusion detection, it appraises network traffic features dynamically, and adopts Accumulator Model to calculate positive and negative features respectively to improve accuracy. We make use of information gain theory to separate continuous features, and construct a decision tree according to the order of feature computed, weighting the feature by its importance. In the end, we give an experiment result and prove that: the model is available.