Mining Frequent Patterns from Network Data Flow

The main objective of network monitoring is to discover the event patterns that happen frequently. In this paper, we have intensively studied the techniques used to mine frequent patterns from network data flow. We devel-oped a powerful class of algorithms to deal with a series of problems when min-ing frequent patterns from network data flow. We experimentally evaluate our algorithms on real datasets collected from the campus network of Peking Uni-versity. The experimental results show these algorithms are efficient.