基于Boosting半监督的网络安全入侵检测算法

Intrusion Detection of Network Security Based on Semi-Supervision

针对网络安全入侵行为升级快、隐蔽性强和随机性高等严重的安全问题,提出了一种基于半监督的网络安全入侵检测算法.该算法利用Boosting建立入侵检测模糊分类器,采用遗传算法进行迭代训练,生成最终的网络安全入侵检测模型.仿真结果表明,该算法有效提高了网络安全入侵检测的性能和效率.与SVM等先进的入侵检测方法相比,该算法能更加准确有效地检测各种类型的入侵,具有良好的检测效果和应用价值.

Stochastic subspace identification was utilized to extract modal parameters from the structural response signals,by which the modal flexibility curvature （MFC） was constructed to detect structural damage. A numerical model of simple beam using line elements was employed to compare the results between the displacement modes and strain modes by means of the above procedure. In addition,a simple beam model constructed by solid elements was also used to investigate the effect of recorded locations of strain signals on the damage identification. The results show that the strain modes are better than displacement modes in locating damage under the effect of noise;however, the recorded places of strain signals should be near the damaged parts.