混合型数据库中入侵检测技术仿真

Simulation of Intrusion Detection Technology for Hybrid Database

在对混合型数据库的入侵进行准确检测,保障数据库安全问题的研究中,由于数据库中的数据属性差异很大,无法形成可供检测的统一阀值特征。传统的检测算法以对数据进行分类后,设置多阀值的形式完成检测,但是,分类过程无法做到完全精确,一旦数据库中数据属性数目过多,不同属性数据掺杂,很难针对不同属性数据进行准确的入侵检测。提出基于粗糙集和信息熵融合算法的混合型数据库中入侵检测方法。利用粗糙集表达复杂信息的能力,构建混合型数据库中入侵特征数据集合,针对该集合中的元素,计算不同特征的信息熵,得到不同入侵特征在入侵检测过程中的重要性,获取简化处理后的入侵特征集合,实现混合型数据库的入侵检测。实验结果表明,利用改进算法进行混合型数据库中入侵检测,能够简化入侵训练集合,缩短检测时间,极大的提高了检测的准确性。

An intrusion detection method for hybrid database is presented based on rough set and information entropy fusion algorithm. By using the rough set＇s ability to express complex information, a set of intrusion characteristics for hybrid database is built, the information entropy of different characteristics is calculated for the elements in the set, the different importances of different invasion characteristics in the process of intrusion detection are got, the set of invasion characteristics after simplified processing is obtained, and the hybrid database intrusion detection is realized. Experimental results show that the improved algorithm can simplify the invasion of the training set, shorten the testing time, greatly improve the accuracy of the detection.