中文摘要：

为提高网络数据的检测效率，将差分进化算法与支持向量机算法融合（DE—SVM）应用到网络入侵检测中。引入自适应算子优化差分进化算法中的交叉概率CR和摄动比例因子F，采用优化的DE算法对支持向量机的参数进行选择，构建DE—SVM入侵检测算法。KDDCUP99数据集的测试结果表明，融合算法提高了网络入侵检测的性能，缩短了训练时间。

英文摘要：

In order to improve the detection efficiency of network data, DE-SVM fusion algorithm is applied to network intrusion detection. The adaptive operators are used to optimize two differential evolution control parameters CR and F. The optimized DE algorithm is applied to solve the SVM parameter selection problem and build DE-SVM intrusion detection algorithm. Simulation experiment is done in KDDCUP 99. It shows that the fusion algorithm improves the performance of network intrusion detection and shortens the training time.