摘要
大量的冗余和噪音数据混合于网络入侵数据中,从而影响到检测的性能和响应。因此,提出基于Fisher-FCBF算法。通过对特征的Fisher分值排序,再使用FCBF算法去冗余,结合SVM,建立分类特征模型,在不降低准确率的前提下,选出最优特征子集,结果表明所提出的方法能够在保证分类准确率的情况下,降低至少11%-21%的计算时间。
A large amount of redundancy and noise data are mixed in the network intrusion data, thus affects the performance and re- sponse of the detection. By sorting the Fisher scores of the feature, uses the FCBF algorithm to reduce the redundancy and us- es SVM to establish the classification feature model. The optimal feature subset is selected without reducing the accuracy. The results show that the proposed method can reduce at least 11% -21% of the calculation time in the case of classification accuracy to ensure.
作者
王浩
石研
WANG Hao SHI Yan(School of Information Science and Technology, Xinjiang University, Urumqi 830046 School of Software, Xinjiang University, Urumqi 830008)
出处
《现代计算机》
2017年第10期7-12,共6页
Modern Computer
基金
国家自然科学基金项目(No.61163052
No.61303231
No.61433012)
国家自然科学基金联合基金项目(No.U1435215)
