大规模数据库高危攻击数据实时挖掘仿真研究  

Real-Time Mining Simulation of High Risk Attack Data of Large-Scale Database

作　　者：李浩 汤哲君 LI Hao1 , TANG Zhe - jun2(1. Academic Affairs Office, Ningxia University, Yinchuan Ningxia 750021, China; 2. Institute of Physics and Electrical Engineering, Ningxia University, Yinchuan Ningxia 750021, China)

机构地区：[1]宁夏大学教务处,宁夏银川750021 [2]宁夏大学物理与电子电气工程学院,宁夏银川750021

出　　处：《计算机仿真》2018年第10期381-384,共4页Computer Simulation

摘　　要：对大规模数据库的高危攻击数据进行挖掘,能有效提高数据挖掘的精度,提高数据库防攻击的性能。当前利用关联规则的映射挖掘算法,对攻击数据进行挖掘时,由于数据较多,数据挖掘的准确度较低,降低了高危数据挖掘的精度。提出基于粒子群优化的攻击数据检测的算法。利用粗糙集的理论对大规模数据库高危攻击的数据进行属性的约简,提高攻击数据属性的依赖度,利用粒子群优化检测算法对大规模数据库高危的攻击数据进行检测,针对粒子群算法存在局部的早熟收敛的问题,采用改进粒子的属性,对粒子群算法进行改进,增加粒子的多样性,经过种群的初始化以速度与位置进行的更新,对粒子适应度的值进行计算,对粒子全局的极值进行更新,进行粒子循环的迭代,得出最优的解,完成对大规模数据库的高危攻击的数据实时的挖掘。实验的结果表明,利用所提的算法,在减少内存的占用容量的同时,有效地提高了数据实时挖掘的精度。An algorithm for detecting attack data based on particle swarm optimization was put forward. Firstly, this research used the theory of rough set to carry out attribute reduction of high - risk attack data in large - scale database and improve the degree of dependence for attack data attribute. Then, our research used particle swarm optimization algorithm to detect high - risk attack data in large - scale database. For problem of local premature convergence of particle swarm optimization, the research used the attribute of improved particle to improve the particle swarm algorithm and increase the diversity of particles. Through the initialization of population and the updating of speed and position, the research calculated particle fitness values and update global extreme value of particle. Finally, the iteration of particle recycling was carried out to find the optimal solution. Thus, we completed the real - time mining for high - risk attack data in large - scale database. Simulation results prove that the proposed algorithm can improve the accuracy of real - time data mining when reducing the amount of memory footprint.

关 键 词：大规模数据库 高危攻击数据 数据挖掘 粒子群优化检测 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]