Intelligent Compliance Checking System for BIM Model Based on Semantic Similarity Matching
This study proposed a web-based intelligent inspection system for BIM models based on semantic similarity matching.The application of machine learning-based natural language processing technology enables automated extraction and analysis of semantic information from BIM models.This information is then matched with a normative knowledge graph ontology expressed based on the IFC standards.A combination of two similarity matching metrics,namely,Levenshtein distance and Cosine Similarity,was employed to perform the similarity matching inspection of the BIM model's semantic information.Additionally,a comparative evaluation was conducted between precise and similarity matching algorithms,the results of which indicate that similarity matching outperforms precise matching in overall inspection performance and enhances matching accuracy.The developed intelligent inspection system demonstrates substantial versatility and adaptability,capable of supporting the inspection of models following diverse industry standards and extending the application range of BIM model matching research.The findings of this study are expected to facilitate the more efficient application of BIM technology in the building industry,thereby elevating the management quality and efficiency of engineering projects.
BIMcompliance checkingnatural language processingontologysemantic mappingsimilarity matching
NETL
NSTL
万方数据
