基于当前多分类器融合的入侵检测系统检测性能、训练时间和检测时间等问题，本文提出了利用主成分分析和模糊综合的入侵检测方法。该方法利用主成分分析在特征提取上的优势，首先利用基于总体协方差的PCA和基于总类内散布矩阵的PCA的两种主成分分析方法提取入侵特征；然后通过KNN分类器给出初步的识别结果，依据两个分类器存在的互补性，将此识别结果采用模糊综合方法进行融合，得出最终识别结果。利用KDD CUP’99的标准入侵检测数据集进行实验，实验结果验证了该方法的有效性。该方法提高了入侵检测的整体性能,降低了训练时间和检测时间。

In order to solve the detection performance,training time and detection time,an intrusion detection method using fuzzy integration based on two different Principal Component Analyses（PCA）feature analyses is presented. Firstly, two different PCAs is applied to network intrusion feature extraction. Then, an initial intrusion detection result is done by two KNN classifiers. The two classifiers can overcome the shortcomings of each other.The last step is to form the final result by fusing these results using fuzzy integration. Experiments have been done on the datasets in KDD99,and the results show that the method is effective and the intrusion detection’s total performance is improved.