Abstract: This paper proposed a new Support Vector Machine (SVM) for anomaly intrusion detection method based on Latent Semantic Indexing (LSI). In this paper, the PARPA’98 data sets were chosen as training and testing data sets, experiments showed that our method had a higher detection rate and a lower false positive rate, and can greatly reduce the computation complexity. It is an effective anomaly identifying and detecting method

Key words: Intrusion Detection, Support Vector Machine, Latent Semantic Model

摘要： 本文提出了一种基于潜在语义索引（LSI）和支持向量机（SVM）的异常入侵检测方法。选取PARPA′98 BSM数据集作为训练数据和测试数据，通过实验比较和分析表明：基于LSI和SVM方法的入侵检测系统具有较高的检测率和较低的虚警率，且能大大减低计算的复杂性，是一种有效的异常识别和检测方法。

关键词: 入侵检测, 支持向量机, 潜在语义索引