Although studies on conceptual metaphor and metonymy are being actively conducted in the field of metaphor identification, they do not secure objectivity because they are mainly conducted in the way of generating and presenting data based on the resea ...
Although studies on conceptual metaphor and metonymy are being actively conducted in the field of metaphor identification, they do not secure objectivity because they are mainly conducted in the way of generating and presenting data based on the researcher's intuition. Conceptual metaphor or conceptual metonymy identification extracts from the corpus and checks the frequency of words or uses syntactic patterns of conceptual metaphors, so it is very difficult to find related vocabulary, parts of speech, and syntax. In addition, identifying a conceptual metaphor is difficult because it is based on a single vocabulary or uses a single part of speech or syntactic pattern.
The final goal of this study was to collect and analyze metaphor big data, build a metaphor knowledge base, analyze semantic contextual information from the constructed metaphor knowledge base, and create a metaphorical expression method and automatic script. Through this study, we tried to present a methodology that can be used in various fields of linguistics by constructing a metaphoric knowledge base with high utilization value beyond the corpus of a very limited scale. As for the progress of the study, first, a method of automatically generating and expanding the metaphor knowledge base was studied through semantic lexical relationship analysis, structural analysis of sentences, and metaphor pattern extraction. Basic research was conducted to extract grammatical characteristics of conceptual metaphors and conceptual vocabularies of source and target areas through the analysis of existing metaphor data sets.
Second, conceptual metaphor identification proceeded by sentence unit, grammatical pattern, and word level. As for the grammatical patterns used in conceptual metaphors, verbs and adjectives tend to be associated with nouns, and conceptual metaphorical expressions were identified by defining the relationship between verbs and nouns and adjectives and nouns as the main patterns.
Third, the Hearst Pattern was expanded to define the pattern for identifying the upper-lower relationship, and a study was conducted to extract the semantic relationship between the source area and the target area in the conceptual metaphor text. In addition, knowledge graph technology was applied to expand the metaphor knowledge base. In order to expand the knowledge base extracted from metaphorical text, based on the extracted entities and relationships, a method for expanding the knowledge base using Wikipedia and Google news was studied, and a method for visualizing it was implemented.