DGA Domain Name Segmentation Method Based on Natural Language Word Frequency Weighting

In order to improve the survival rate of botnets, attackers use Domain Generation Algorithm to rapidly generate a large number of domain names, so quickly and accurately identifying DGA domain names is significant to tracking and detecting botnets. Existing domain name segmentation methods do not consider the difference between the degree of confusion of DGA domain names and that of normal domain names, which can easily cause underreporting of DGA domain names, thus affecting the effectiveness of domain name detection. In this paper, after analyzing the generation characteristics of normal domain names and DGA domain names, we propose a DGA domain name segmentation method based on natural language word frequency weighting, and the experimental results show that the accuracy rate of our method is as high as 99.85%, and the false alarm rate is reduced to 1.16%.