DR & FPR of each intrusion category by the different algorithms.

The result showed that the DR and FPR of PSO-GRNN and AIAE-GRNN were higher than GA-GRNN and FCM. And the DR and FPR of AIAE-GRNN were higher slightly than PSO-GRNN.DR & FPR of each intrusion category by the different algorithms.

研究结果表明，粒子群优化广义回归神经网络（PSO-GRNN）与自适应免疫进化广义回归神经网络（AIAE-GRNN）的检测率（Detection Rate, DR）与误报率（False Positive Rate, FPR）均优于遗传算法优化广义回归神经网络（GA-GRNN）与模糊C均值聚类（FCM）；其中，AIAE-GRNN的检测率与误报率略胜PSO-GRNN一筹。不同算法下各入侵类别的检测率与误报率。