基于镜像随机分离的计算机攻击识别技术研究  

Research on Identification of Computer Attack Based on Random Mirror Separation Technology

作　　者：马蕾[1] 

机构地区：[1]北京电子科技职业学院,北京100016

出　　处：《科技通报》2014年第6期125-127,共3页Bulletin of Science and Technology

摘　　要：传统的计算机识别系统采用普通特征提取的方法,对常规计算机攻击具有较好的拦截性能,但对于高级攻击,拦截性能很差。提出一种基于镜像随机分离的计算机攻击识别技术,首先对访问计算机系统的所有数据进行训练,依据模型分别结果,建立相应的攻击库,同时对攻击的变种进行识别和拦截,由于镜像随机分离具有很好的模糊识别性能,所以对于各种计算机攻击具有很好的拦截性能。采用多种不同的攻击进行测试实验,结果显示,基于镜像随机分离的计算机攻击识别系统能很好的识别多种不同的攻击,可以被广泛应用到识别系统中。In conventional computer identification system, the common feature extraction method was used, it has good inter-ception performance for conventional computer attacks, but for advanced attack intercept, the performance was poor. The identification of computer attack was proposed based on random mirror separation technology, first of all, the data access to computer systems was collected for training, according to the model result, the appropriate attack library was established, and while the attack variants to identify and intercept, the mirror had good fuzzy random separation recognition perfor-mance, so for the variety of computer attack, the result was good. A variety of different attacks was done to do the test experi-ment, and the result shows that with random mirror separation technology, the attack can be identified well, so it can be widely applied in the recognition system.

关 键 词：镜像随机分离 计算机攻击 识别库 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]