Library Trends

Abstract:

Tang poetry, which still attracts a lot of research interest due to its unique value and wide audience, is one of the highest achievements of Chinese culture. Tang poetry data have the characteristics of multisource heterogeneity and complex association. In this paper, we construct a knowledge service platform, KnowPoetry, based on a domain-specific knowledge graph of Tang poetry to support data-driven research in the field of Tang poetry. We first acquire knowledge service requirements from Tang poetry researchers. Then, we construct domain ontology and propose a human-machine coordinated knowledge extraction framework. The framework extracts and merges entities of poetry, poet and poet experience, and multidimension relations among them from massive Tang poetry data. Finally, we construct a large-scale domain-specific knowledge graph. Based on the knowledge graph, KnowPoetry provides researchers with Tang poetry domain knowledge services including knowledge navigation, association analysis, and semantic query. This platform can support researchers in navigating and discovering the inner relationship between Tang poetry in terms of emotion and imagery, analyzing poet social circle and life experiences, and realizing multidimensional semantic query. In the experiment, we evaluate the effect of domain knowledge service in supporting Tang poetry research.