摘要
为了进一步提高入侵检测的准确率,提出了一种融合半监督LDA和PSO-SVM方法,使用累计贡献率ω确定主成分分析法(PCA)占半监督LDA算法比例,然后使用PSO参数寻优算法对支持向量机进行参数寻优,最终得到入侵检测模型。实验结果显示,与单一的PCA或LDA与PSO-SVM组合相比,这种半监督LDA和PSO-SVM方法具有优势,对异常行为的查准率比PCA或LDA与PSO-SVM方法组合准确率较高。
In order to further improve the accuracy of intrusion detection,a semi-supervised LDA and PSO-SVM method is proposed.The cumulative contribution rate is used to determine the proportion of principal component analysis(PCA)to semi-su⁃pervised LDA algorithm.The PSO parameter optimization algorithm is used to optimize the parameters of support vector machines,and finally the intrusion detection model is obtained.The experimental results show that this semi-supervised LDA and PSO-SVM method has advantages over the single PCA or LDA and PSO-SVM combination,and the accuracy of abnormal behavior is higher than that of PCA or LDA and PSO-SVM combination.
作者
詹琉
ZHAN Liu(Guangdong University of Technology,Guangzhou 511400)
出处
《计算机与数字工程》
2021年第1期158-162,共5页
Computer & Digital Engineering
