网络入侵检测中属性分组的随机森林算法

入侵检测是数据挖掘的一个重要应用领域,目前基于数据挖掘的入侵检测方法很多,而基于随机森林的方法具有比较好的性能,但仍存在一些问题。通过分析网络入侵数据得到不同输入属性与分类结果的关系,提出了一种基于属性分组的随机森林算法,并应用该算法对KDD’99数据集分类。实验结果表明,该算法的训练速度和分类准确率都比原算法有较大提高。

Random Forests Algorithm with Feature Grouping in Network Intrusion Detection

Intrusion detection is one of the important application areas of data mining. At present, there are many approaches of intrusion detection based on data mining. Although random forests method has shown better performance than some other methods, but it still has some problems. After analyzing the network intrusion data set we get the relationship between the different input features and the result of classification, so we propose a new random forests algorithm based on feature grouping, and then we applied it in KDD＇ 99 data set. The test result of our new algorithm show that this method is much better than before in accuracy and speed.