【摘要】 用于异常检测的机器学习方法,如神经网络和支持向量机,都对训练样本的噪声非常敏感,进而导致推广能力和分类准确性的下降。为了解决上述问题,论文提出一种新的基于健壮支持向量机的方法。先将RSVM与标准SVM作了对比,然后使用1998DARPABSM的数据作为评估数据。实验表明,该方法在入侵检测的准确率、误检率和有噪声情况下的推广能力和运行时等多项指标上都有良好的表现。更多还原

【Abstract】 The machine learning techniques used for intrusion detection,including neural network and support vector machine,are sensitive to noise of training samples,and lead to the poor generalization ability and classification accuracy.In this paper,we propose a new support vector machine based on Robust Support Vector Machine for Anomaly Detec-tion using the1998DARPA BSM data set as evaluation data.The performance of Robust Support Vector Machine(RSVM)has been compared with that of standard Support Vector Machines.The results indicate the superiority of RSVM not only in terms of high intrusion detection accuracy and low false positives but also in terms of their generalization ability in the presence of noise and running time.更多还原