摘要
为了提高入侵检测系统的检测精度和效率,提出一种基于改进鸡群算法(ICSO)和核极限学习机(KELM)的入侵检测模型(ICSO-KELM)。考虑到模型中特征选择与分类器参数优化的相互影响,利用具有较好全局优化能力的改进鸡群优化算法优化核极限学习机正规化系数和核函数参数的同时选择最优特征子集。仿真实验结果表明,该方法有效地适配了入侵检测中的特征选择和分类器参数,与SVM、KNN等方法相比,其检测准确率和效率有明显提升,误报率也有所降低。
In order to improve the detection accuracy and efficiency of intrusion detection system,an intrusion detection model(ICSO-KELM)based on improved chicken colony algorithm(ICSO)and kernel limit learning machine(KELM)is proposed.Considering the interaction between feature selection and classifier parameter optimization in the model,the improved chicken colony optimization algorithm with better global optimization ability is used to optimize the normalization coefficient and kernel function parameters of the kernel limit learning machine and select the optimal feature subset at the same time.The simulation results show that the method adapts the feature selection and classifier parameters effectively.Compared with SVM,KNN and other methods,the detection accuracy and efficiency are improved obviously,and the false alarm rate is also reduced.
作者
谭敏生
蔡畅
TAN Minsheng;CAI Chang(Nanhua University,Hengyang Hunan 421000,China)
出处
《自动化与仪器仪表》
2020年第12期1-5,共5页
Automation & Instrumentation
基金
湖南省自然科学基金项目(No.2017JJ4048)
湖南省教育厅科学研究重点项目(No.18A230)
湖南省财政厅科学研究项目(No.20191550502)
2019湖南省研究生科研创新项目(No.CX20190736)。
关键词
鸡群优化算法
核极限学习机
特征选择
分类器参数优化
入侵检测
chicken swarm optimization algorithm
kernel extreme learning machine
feature selection
classifier parameters
intrusion detection
