论文部分内容阅读
由于入侵检测的数据都是海量高维数据,提出一种基于主成分分析(PCA)的特征提取方法,以提高入侵检测的处理效率。选用Kddcup’99网络连接数据集进行预处理和PCA特征提取后,分别通过BP神经网络和Kohonen神经网络进行训练和测试,分析检测率,误报率,训练时间和检测时间。实验结果表明,基于PCA的BP神经网络能减小入侵检测的运算量,提高入侵检测的识别效果。
Because the intrusion detection data are massive high-dimensional data, a feature extraction method based on principal component analysis (PCA) is proposed to improve the processing efficiency of intrusion detection. After preprocessing and PCA feature extraction using Kddcup’99 network connection dataset, they were trained and tested by BP neural network and Kohonen neural network respectively. The detection rate, false alarm rate, training time and detection time were analyzed. The experimental results show that BP neural network based on PCA can reduce the computational complexity of intrusion detection and improve the recognition effect of intrusion detection.