Construction of bilingual knowledge graph based on meteorological simulation

Geographical simulation is the core function of virtual geographic environments (VGEs). Providing knowledge support has become a key issue in current research. The literature involves massive knowledge, especially for meteorological simulation, an important component of the geographic environment. However, this knowledge is noisy, unstructured, and difficult to share and reuse. Based on weather research and forecasting models, this study constructs a cross-language knowledge graph to manage difficult-to-handle knowledge and to make complex knowledge structured and more complete. First, we propose a cross-language knowledge graph construction framework. Second, the bidirectional long short-term memory-conditional random fields and translation-based multilingual knowledge graph embedding model are used to extract knowledge and align cross-lingual entities. Finally, the bilingual knowledge graph is visually stored with a graph database. The framework and method can provide efficient queries and analyses for the meteorological field, while helping to provide a knowledge base for knowledge-driven VGEs.