基于神经网络的网络入侵检测系统实现
文献综述
摘要：计算机网络安全问题是随着网络、特别是Internet的发展而产生的，受到国家的普遍重视。计算机网络的连通性和开放性给资源共享和通信交流带来了巨大的便利，但也使网络安全的问题愈发严重，开放性使入侵者可以预知系统的行为。如今最常用的网络安全技术就是防火墙。但防火墙只是防御为主，通过防火墙的数据便不再进行任何操作。黑客利用多种入侵手段攻击目标主机,通过防火墙拦截,安全访问控制,对数据加密处理以及身份认证等传统的网络安全防御技术也无法感知,静态的防御体系已经无法满足当前的安全需要。为了应对这一系列问题，入侵检测技术应运而生。及时的检测入侵以及作出相应对网络安全有着积极的意义。入侵检测系统可以进行实时的检测，发现入侵行为即可做出反应，是对防火墙弱点的修补。本文运用神经网络的概念来解决入侵检测系统中出现的若干问题，设计一个基于神经网络的入侵检测系统，自动识别网络入侵现象，并进行系统预警和记录。
关键词： 神经网络 入侵检测系统 python
THEME：Implementation of network intrusion detection system based on Neural Network
Abstract: With the development of network, especially Internet, the problem of computer network security comes into being, which attracts a lot of attention. The connectivity and openness of computer network have brought great convenience to resource sharing and communication, but it also makes the problem of network security more and more serious. The openness enables the intruder to predict the behavior of the system. Nowadays, the most commonly used network security technology is firewall. But the firewall is only defense oriented, and the data passing through the firewall will not be operated any more. Hackers use a variety of intrusion means to attack the target host, through the firewall interception, security access control, data encryption and identity authentication and other traditional network security defense technology can not be perceived, static defense system has been unable to meet the current security needs. In order to deal with this series of problems, intrusion detection technology appeard. Timely detection of intrusion and the corresponding response to network security has a positive significance. Intrusion detection system can carry out timely detection. If it finds the intrusion behavior, it can react, which is the repair of firewall weakness. This paper involves the concept of neural network to solve some problems in the intrusion detection system and designs an intrusion detection system based on neural network, which can automatically identifies the network intrusion phenomenon for early warning and recording.
Keywords： Neural Network, Intrusion Detection System, Python
前言
入侵检测技术自20世纪80年代早期提出以后，经过20多年的不断发展，从最初的一种有价值的研究想法和单纯的理论模型，迅速发展出多种多样的各种实际原型系统，并且在近10年内涌现出许多商用入侵检测系统产品，成为计算机安全防护领域内不可或缺的一种重要的安全防护技术。随着时代进步，不少研究人员又提出了不少新的检测算法，这些检测方法从不同的技术角度来看待入侵检测的基本问题，并利用了许多人工智能和机器学习的算法，试图解决虚假警报等问题。本文就是基于人工神经网络来实现一个入侵检测系统，这种检测技术具有更大的发展前景。
研究的目的与意义
