Prediction Of Network Security Situation Based On Cloud

The index system of network security situation factors ought to reflect the network security situation accurately and entirely, and there isn't an index system standard issued now. The coverage of some existing index systems is narrow. Situation prediction is an indispensable component of network security situation awareness system. The existing algorithms of network security situation prediction depend on the initial training data, and the objectivity of predicting results is insufficient. These issues are studied in this thesis and the thought and method of network security situation prediction based on cloud is proposed, and the objectivity of predicting results is improved.The main tasks and contributions of this thesis are as follows.(1) An index system of network security situation factors was proposed to provide comprehensive and reliable sustainment for situation assessment and prediction, and the methods of obtaining situation factors data and calculating index were presented.(2) A prediction method combined with quantitative and qualitative of network security situation based on cloud was proposed, and the relationship of quantitative value and qualitative concept was established. The future situation was predicted by combining the current trend with the evolvement rules mining from history data. The objectivity and intelligibility of predicting results were improved for the method didn't need the initial data training.(3) A prediction rules mining algorithm of network security situation based on cloud was proposed. A soft partition method of attributes interval based on cloud model was adopted to solve the problems of that the internal relationship of boundary elements lost by hard partition during the exchanging of quantitative and qualitative. The concepts of "support cloud" and "confidence cloud" and the algorithm of minimum support and confidence thresholds based on cloud were proposed to solve the problem of that the selection of minimum support and confidence thresholds was influenced by subjective factor in prediction rules mining. The objectivity of predicting results was improved.(4) The algorithms were designed and implemented. The experimental results show that the algorithm of network security situation prediction based on cloud is feasible. The situation prediction results are accurate, and the objectivity and accuracy of predicting results was improved.The research work in this thesis increases the objectivity of network security situation prediction. The timely, accurate and forward-looking information of network security situation is provided for the manager in decision making of network security defense. It provides a new way for network security situation awareness.