An Identification Algorithm of Attacking Programs based on Quadratic Feature Selection and Fast Decision Tree
Dengzhou Shi, Saihua Cai, Songling Qin, Zhenxin Wang, Qiyong Zhong, Ling Zhou · 2021 IEEE 21st International Conference on Software Quality, Reliability and Security Companion (QRS-C) · 2021
With the rapid development of Internet, the scale of the network is gradually expanding, which brings convenience to our life, as well as brings many network security issues, such as network attack, network intrusion, etc. Therefore, how to identify attack programs in the network traffic has become one of the important research directions in the field of network security. At present, machine learning methods can be used to identify network attack programs, but the existing feature selection methods have the disadvantage of not being able to effectively extract key features, and the classification model also has the problem of low efficiency. In this paper, we propose an attack programs identification algorithm based on Quadratic Feature Selection and Fast Decision Tree (QFS-FDT). Firstly, the data set is sampled and normalized in layers to obtain a better quality. And then, the key feature subsets are extracted by quadratic feature selection based on principal component analysis and information gain. Finally, the classification model is trained by the fast decision tree algorithm based on the key feature subset to identify the attacking programs. Extensive experimental results show that the proposed method can accurately identify the attacking programs with an improvement of 3% to 6%.