Research Article
Detecting Dictionary Based AGDs Based on Community Detection
@INPROCEEDINGS{10.1007/978-3-030-63086-7_3, author={Qianying Shen and Futai Zou}, title={Detecting Dictionary Based AGDs Based on Community Detection}, proceedings={Security and Privacy in Communication Networks. 16th EAI International Conference, SecureComm 2020, Washington, DC, USA, October 21-23, 2020, Proceedings, Part I}, proceedings_a={SECURECOMM}, year={2020}, month={12}, keywords={Algorithmically generated domains Community detection Machine learning}, doi={10.1007/978-3-030-63086-7_3} }- Qianying Shen
Futai Zou
Year: 2020
Detecting Dictionary Based AGDs Based on Community Detection
SECURECOMM
Springer
DOI: 10.1007/978-3-030-63086-7_3
Abstract
Domain generation algorithms (DGA) are widely used by malware families to realize remote control. Researchers have tried to adopt deep learning methods to detect algorithmically generated domains (AGD) automatically based on only domain strings alone. Usually, such methods analyze the structure and semantic features of domain strings since simple AGDs show great difference in these two aspects. Among various types of AGDs, dictionary-based AGDs are unique for its semantic similarity to normal domains, which makes such detections based on only domain strings difficult. In this paper, we observe that the relationship between domains generated based on a same dictionary shows graphical features. We focus on the detection of dictionary-based AGDs and proposes Word-Map which is based on community detection algorithm to detect dictionary-based AGDs. Word-Map achieved an accuracy above 98.5% and recall rate above 99.0% on testing sets.
