基于系统调用的神经网络异常检测技术

A Neural Network Anormaly Intrusion Detection Technology Based on System Calls

给出了一种基于程序行为的神经网络异常检测技术，通过使用ＥＬＭＡＮ网络，利用了程序执行时的系统调用的周期重复特性，使该系统具有检测新型入侵攻击和具有较低虚警率的特点。

This paper presents an anomaly detection technique for profiling program. By using an Elman network for recognizing recurrent features in program execution traces, system is given the character of ability to detect novel intrusion and of low false positive.