| With the rapid popularization and application,the Internet has changed the entire world,and promote society rapid progress and development.But at the same time due to the characteristics of free and open,which makes the network harm behavior becomes more and more widespread.And caused great harm to the society and the people.How to accurate and comprehensive perception of the current health of the network environment,timely found aggressive behavior,and scientific intelligently for processing which is an important focus of research in the field of network security.Network security situation awareness technology through the integration of large-scale network security equipment collection and analysis security data to achieve a comprehensive assessment of network security situation.To achieve real-time dynamic discover and perceive network security threats,by alerting administrators to take timely measures to enhance information security network security situation.This paper analyzes the key technologies and research results of the security situation system,for security posture prediction,using machine learning methods to predict the security posture time series.Proposed network security situation prediction model based on FOA-SVM algorithm,the method combines excellent SVR algorithm nonlinear fitting capability,and FOA good global optimization capability.Use FOA optimize security posture time series embedding dimension n,SVR penalty coefficient C and RBF kernel function parameters g,thus avoiding the parameters of blindness,improve the network security situation prediction accuracy.And through experiments constructing RBF neural network and PSO-SVM algorithm make comparative analysis,and demonstrate excellent characteristics FOA-SVR algorithm.At the same time design network security situation analysis system architecture.Respectively,data collection and normalization,dynamic correlation,trend calculation and prediction of their detailed design of each module,and to develop a clear process flow giving final realization,and apply the FOA-SVM algorithm in prediction module of the security posture system. |
