Search and Inference of Collision-Preventing Regulations Based on Chinese Semantic Similarity

The 1972 International Regulations for Preventing Collisions at Sea (hereinafter referred to as the “regulations”) is a legal norm used to guide collision avoidance actions to avoid collision accidents. In order to reduce the difficulty of regulation learning, realize the automatic inference of regulations, and finally assist the automatic control of ships, this paper proposes a search and inference method based on sentence similarity, which comprehensively considers the similarity of text semantic and syntactic structure. It is used for semantic search of regulations, knowledge search, and regulation inference. The search result of this algorithm is more accurate and has higher search performance than the traditional algorithm which only considers semantic information.