近年来,网络应用识别在学术和应用领域备受关注和快速发展,已形成一个相对独立的研究领域。基于机器学习的识别方法更是成为研究热点,并且取得一系列初步成果。研究了这方面的相关问题,分别从识别粒度、特征选择及识别算法等方面进行介绍、归纳,并对典型方法进行重点分析,最后指出了存在的问题及未来研究方向。
In recent years, identifying and categorizing network traffic by application type attracts great interests both in the fields of academic and application, and has already become a relatively independent research realm. Furthermore, the application identification approaches based on Machine learning have been hotspots and have been obtained promising preliminary results. This paper surveyed the current machine learning algorithms about application classification, respec tively from fine-grained identification, feature selection, and recognition steps and so on. It focused on the analysis of typical methods and suggested some future research directions.