Network Situation Assessment Based on RST

As the core of Cyberspace Situational Awareness whichis the the future development direction of network management, Network Situation Assessment (NSA) can integrate unit network managements, alter the current complexion that each function unit works in an independent state, provide the comprehensive macroscopic view of network operation status, strengthen the comprehension and control of network, and reducethe burden on administrators. This paper focuses on the characteristics and requirements of network and introduces Rough Set Theory (RST) which has the superiority in machine learning, dealing with massive redundancy information, feature selection, etc. After giving the formal definition of NSA knowledge system, we establish NSA method based on RST, and then discuss data discretization, situation factor selection, assessment rules design and other techniques in detail. The experiment results on real data set NLANR demonstrate the efficiency, effectiveness and scalability.