基于模糊数据分离的网络疑似攻击检测技术  

Network Suspected Attack Detection Technology Based on Fuzzy Data Separation

作　　者：马蕾[1,2] 杜辉[1] 康英健[1] 

机构地区：[1]北京电子科技职业学院电信工程学院,北京100015 [2]北京航空航天大学计算机学院,北京100191

出　　处：《计算机仿真》2016年第3期231-234,共4页Computer Simulation

摘　　要：在对网络疑似攻击检测技术研究过程中,由于网络中的大量疑似攻击,可造成网络崩溃。采用当前的算法进行疑似攻击检测时,攻击特征存在一定的模糊性,需要预先设定参数才能判断出疑似攻击的数量,对一些疑似攻击无法准确分类,存在检测误差较大的问题。提出采用聚类算法的模糊数据分离的网络疑似攻击检测方法。上述方法先利用聚类算法对疑似攻击中的网络连接数据进行模糊数据分离,组建对检测有价值的疑似攻击的数据集合,在组建的集合中利用模糊数据分离的联合评分偏离度判断出疑似攻击网络数据,进而精确的的完成了模糊数据分离的网络疑似攻击检测。仿真结果证明,提出的聚类算法的模糊数据分离网络疑似攻击检测方法精确度高,适用性强。Due to the large number of suspected attacks in network, the current attack detection algorithm needs to preset parameter to determine the number of suspected attacks, results in large detection error. A suspected attack detection method is presented based on fuzzy data separation and clustering algorithm. The method uses clustering al- gorithm to classify the network connection data in suspected attacks, a data set of suspected attacks which are valua- ble for detection is built, the suspected attack network data are judged according to the joint score deviation of the fuzzy data separation, and the network suspected attack detection is accurately completed based on fuzzy data separa- tion. Simulation results show that the proposed network suspected attack detection method has high precision and wide applicability.

关 键 词：聚类 疑似攻击 目标项目 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]