摘要： 为实现海量网络数据的入侵检测，将支持向量机应用于入侵检测中。在入侵检测实验中，通过数据筛选策略，减少建立检测模型所需要的样本数，根据每个特征属性的重要性赋予不同权重，设计有特征加权的支持向量机算法。实验结果表明，该算法能缩短检测模型的建立时间，提高检测精度，降低漏报率。

关键词: 支持向量机, 入侵检测, 数据筛选, 特征加权

Abstract: Aiming at the problem that a huge mass of network data are real-time processed, this paper proposes application of support vector machine in intrusion detection. The number of samples for building detection model is decreased by applying data filtering strategy on the intrusion detection experiments. According to the importance measurement of each feature attribute, different weightings are given. It designs the feature weighted Support Vector Machine(SVM) algorithm. Experimental results demonstrate that the algorithm can effectively shorten building detection model time and improve detection accuracy. It can also lower false negative rate.

Key words: Support Vector Machine(SVM), intrusion detection, data filtering, feature weighted

中图分类号: 

• TP393