Terminology Semantic Sememe Tree Knowledge Graph

Knowledge graphs are foundational language resources and research tools derived from corpus studies. Existing graphs mostly focus on factual or event-based descriptions. Semantic category graphs for domain-specific terminology are scarce. Existing terminology repositories mostly remain at the level where signifiers and signifieds are not fully separated, lacking independent modeling and semantic linking. The Terminology Semantic Sememe Tree Knowledge Graph proposes a representation method for terminological concepts, which separately encodes signifiers (terms) and signifieds (concepts), based on a sememe tree structure. Our knowledge graph consists of three parts: the sememe system repository, the term record repository, and the relation repository. It covers core elements including terms, concepts, concept relations, sememes, and dynamic roles. Unlike other terminological semantic knowledge graphs that integrate terms and definitions into unified descriptions, this knowledge graph embeds such relationships within the sememe tree structures, which facilitates the implementation of semantic computation.