摘要： 从编码角度出发，应用稀疏理论学习鲁棒特征。在训练过程中，融合监督类别信息，采用判别式K-SVD算法，优化学习过完备字典和线性判别函数。在测试过程中，将稀疏编码系数作为数据的表示形式，以增强表示力和判别力。实验结果表明，基于判别式K-SVD的方法能获得较高的检测率，且误报率较低，对不平衡数据集也有较好的鲁棒性。

关键词: 稀疏表示, 奇异值分解, 支持向量机, 协同, 入侵检测

Abstract: This paper applies the theory of sparse representation to learn robust features. By fusing class information into the training process, it applies discriminant K-SVD algorithm to intrusion detection, which optimizes the over-completed dictionary and the linear discriminant function together. When testing, it uses the sparse coefficients as sample’s feature, which is more effective representational and discriminative power. Experimental results demonstrate that it can guarantee higher detection rate and lower false alarm rate. Meanwhile, it has good robustness in the imbalanced dataset experiment.

Key words: sparse representation, Singular Value Decomposition(SVD), Support Vector Machine(SVM), cooperation, intrusion detection

中图分类号: 

• TP311