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 current large database, widely exist anti-jamming poor, largely due to the huge amount of testing data makes traditional algorithm in local search the low efficiency of defects, causing the potential risk to 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 multiple calibration, the largest regional risk data attributes are calculatied. Through detailed virtual area, avoid traditional exhaustively algorithm, there is the defects 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 detection accuracy, good results.
展开▼