Construction of Meteorological Simulation Knowledge Graph Based on Deep Learning Method

With the maturity of meteorological simulation technology, the research literature in this field is undergoing a rapid increase. The published literature can provide useful guidance for current research to get scientific results; however, it tends to be rather time consuming to obtain exact knowledge from massive literature, and it is necessary to transform the literature into structured knowledge to meet the efficient management, sharing, and reuse of meteorological simulation knowledge. In this paper, methods of meteorological simulation knowledge extraction and knowledge graph construction are proposed. A deep learning model based on bilateral long short-term memory-conditional random field (BiLSTM-CRF) is used to extract the meteorological simulation knowledge from the massive literature. Then, the Neo4j graph database is used to construct the meteorological simulation knowledge graph. Based on the meteorological simulation knowledge graph, it can realize the structured storage and integration of meteorological simulation knowledge, which can bridge the gap in the transformation of massive literature to sharable and reusable knowledge. Furthermore, the meteorological simulation knowledge graph can be used as an expert resource and contribute to sustainable guidance and optimization for meteorological simulation research.