近些年, “W annaC ry”等勒索软件网络安全问题层出不穷，对我国的互联网造成不可估量的损害。网络入侵检测 系统作为弥补防火墙防御网络威胁的有效的第二道闸门，扮演着保护计算机的重要角色。首先，介绍入侵检测的定义以及研 究现状；其次，介绍机器学习算法及其在解决网络空间安全问题的一般流程，机器学习在入侵检测中的具体应用，尤其随机森 林算法、贝叶斯算法和其他几种主流机器学习算法在入侵检测中取得的进展；最后，讨论了机器学习算法在入侵检测系统中发 展方向。

In recent years, "WannaCry" and other ransomware network security problems have emerged in an endless stream, causing immeasurable damage to the Internet. The network intrusion detection system plays an important role in protecting the computer as an effective second gate to make up the firewall against network threats. This article firstly introduces the definition and research sta- tus of intrusion detection. Secondly, it introduces the machine learning algorithms and their general processes for solving cyberspace security problems. The specific applications of machine learning in intrusion detection are introduced, especially random forest algo- rithm, Bayesian algorithm and other mainstream machine learning algorithms in the progress of intrusion detection; finally, the de- velopment direction of machine learning algorithms in intrusion detection systems is discussed.