一种基于虚构数据区多重校验的数据库入侵检测模型  

Based on an Imaginary Area Calibration of Multiple Database Intrusion Detection Model

作　　者：孙萍萍[1] 

机构地区：[1]杭州职业技术学院,杭州310018

出　　处：《科学技术与工程》2012年第35期9576-9579,共4页Science Technology and Engineering

基　　金：浙江省高等职业教育研究会课题(YB1103)资助

摘　　要：当前的大型数据库中,广泛存在着抗干扰性差,主要是由于检测数据量的巨大使得传统的算法陷入了局部搜索效率低的缺陷当中,造成对潜在风险检测效果不明显等弊端。提出了一种基于虚构数据区域多重校验的数据库入侵检测模型。通过运用数据的相似风险属性组成一个虚拟的小型风险区域,运用多重校验的方法,计算区域中风险最大的数据属性。通过细化虚拟区域,避免传统穷举算法中存在搜索能力不强的缺陷。对潜在的数据入侵风险进行有效的检测。实验表明,该算法提高了一些大型网络数据库入侵检测的准确率,取得了不错的效果。The cunent large database, widely exist anti-jamming poor, largely due to tile huge amol,nt of tes- ting data makes traditional algorithm in local search the low efficiency of＂ defects, causing the potential risk fo test the effect is not obvious disadvantages, etc. The fictional area calibration of multiple database intrusion detection model is put forward. Through the use of risk attribute data of a virtual small risk area, using the method of multipie calibration, the largest regional risk data attributes are ealculatied. Through detailed virtual area, avoid traditional exhat, stively algorithm, there is the defeets of search capability is not strong, the potential risk for effective data invasion detection. Experiments show that the algorithm enhances some large network database intrusion detec- tion accuracy, good results.

关 键 词：入侵检测 多层网络 特征参数 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]