Recommendation of cyber attack method based on knowledge graph

Abstract

The lack of correlation between the new cyber attack intelligence and the knowledge database makes it difficult for the recommendation system to recommend effective cyber attack knowledge in terms of the new security intelligence. This paper proposes an cyber attack method intelligence recommendation algorithm based on knowledge graph. Firstly, it proposes a cyber attack knowledge graph construction scheme based on four kinds of open security databases. Then, based on the idea of collaborative filtering recommendation, the meta-path is introduced to describe different relations between nodes. At last, the recommendation list could be generated by calculating the correlation score of each path with node vector. Compared with the content-based recommendation method, the effectiveness of this method is proved.