摘要
众所周知,云数据库已在各行各业广泛应用并且发挥着不可替代的作用。但正是云数据库的可扩展、分布式以及虚拟化的特性,导致了云数据库面对恶意行为时表现出明显的脆弱性。因此为了提高云数据库的安全性,开展针对恶意行为智能检测的相关研究是至关重要的。对粒子群优化算法进行了改进并与核极限学习机相结合,提出了一种基于改进粒子群优化和核极限学习机的恶意行为识别模型(KE-VP),通过分析云数据库上的网络流量来检测是否存在恶意行为并将恶意行为准确分类为具体的攻击类型。通过实验分析,发现KE-VP与现有方案相比具备更好的检测精度和检测效率。
As is known,cloud database is widely deployed and playing an irreplaceable role in all sectors.However,due to its features of scalability,distributability and visualizability,cloud database is apparently vulnerable to malicious behaviors.Thus for enhancing the security level of cloud database,it is vital to carry out research on smart detection of malicious behaviors.In this paper,the particle swarm optimization algorithm is first improved.Combining with kernel extreme learning machine,a newly malicious behavior detection paradigm based on improved particle swarm optimization and kernel extreme learning machine(KE-VP)is proposed.By analyzing traffic flow in the cloud database,KE-VP module tries to detect whether malicious behaviors exist.If so,it will classify those behaviors into specific attack type.Via analysis on top of experiments,we observe that KE-VP outperforms existing paradigms in terms of accuracy and efficiency.
作者
李玉玲
朱咏梅
刘东红
顾振飞
LI Yuling;ZHU Yongmei;LIU Donghong;GU Zhenfei(Electronic Technology and Engineering InstitutionꎬShanghai Techincal Institute of Electronics Information,Shanghai 201411,China;School of Network and Communication,Nanjing Vocational College of Information Technology,Nanjing 210023,China)
出处
《电子器件》
CAS
北大核心
2022年第1期1-6,共6页
Chinese Journal of Electron Devices
关键词
恶意行为检测
云数据库
粒子群优化
极限学习机
malicious behavior detection
cloud database
particle swarm optimization
extreme learning machine
