改进果蝇算法优化加权极限学习机的入侵检测

An Intrusion Detection Algorithm Based on IFOA and WELM

提出一种改进的果蝇算法优化加权极限学习机入侵检测算法,利用加权极限学习机训练时间短、泛化性能好等优点,对NSL-KDD入侵检测数据集中的不均衡现象,增加少数类攻击的权重,使对网络攻击中稀有攻击的检测率比传统机器学习方法有大幅提高;用迭代步长自适应调整的果蝇优化算法,对加权极限学习机中的隐含层输入权值和偏置进行全局寻优,以避免算法陷入局部最优解,实现了对NSL-KDD入侵检测数据集的分类。实验表明:本算法对稀有攻击的检测率和分类准确率均有提高,误报率有所降低。

An intrusion detection algorithm of WELM optimized by IFOA is proposed. The advantages of short training time and good generalization performance of WELM are used, and the weight of minority attacks is increased, so that the recall rate of minority attacks in network attacks is greatly improved.The FOA with adaptive adjustment of the iterative step size is used, so the input weights and bias of the hidden layer in the WELM are globally optimized to avoid the algorithm falling into local optimal solution and realize the classification of the NSL-KDD intrusion detection data set. The experimental results show that the proposed algorithm improves the recall rate of minority attacks and the accuracy of the overall classification, and reduces the false positive rate.