Research on Construction and Retrieval Mechanism for Cross Domain Engineering Archives Based on Knowledge Graph

With the development of social economy and the increase of engineering construction projects, a large number of engineering archives of various industries have been constructed. Archives are often scattered across different units and institutions, making it difficult to achieve centralized management and sharing of archives, resulting in prominent problems such as missing and errors while retrieving cross domain engineering archives. Therefore, we propose the Construction and Retrieval Mechanism for Cross Domain Engineering Archives Based on Knowledge Graph (CRM4CDA) for improving retrieval performance. In order to construct a visual knowledge graph for cross domain engineering archives, we combine Protégé tools with Neo4j graph database and integrate fragmented data to build ontology and extract entities for achieving knowledge fusion and integration. In order to improve retrieval performance, we employ semantic analysis, intelligent retrieval, and inference to improve retrieval efficiency. Knowledge association technology and transfer methods are also utilized to promote cross domain retrieval performance. At the same time, various measures are taken to ensure the privacy protection for archive retrieval. Both theoretical analysis and practice have shown that our CRM4CDA mechanism has the characteristics of high retrieval accuracy, recall rate, and efficiency. Further research will be carried out for exploring the deep integration of artificial intelligence and knowledge graphs to further improve the intelligence level of archive management.