This thesis explores the application of Knowledge Graphs (KGs) and Large Language Models (LLMs) to enhance project knowledge management in the construction industry. The research addresses long-standing challenges in knowledge retrieval, organization, and analysis within the sector by proposing a novel framework that leverages the semantic understanding capabilities of LLMs to construct KGs from project communication data. The methodology was tested using an open-source BCF file dataset, representing standardized communication data in BIM projects. The methodology involves a multi-step process: first, setting themes of interest for analysis; second, using LLMs to construct a knowledge graph representing both explicit and implied content from BCF comments; and finally, exploring various querying methods to gain insights from the dataset. The framework demonstrates the ability to transform unstructured project communication data into a structured, queryable format, enabling detailed post-project analysis at a scale previously unfeasible with manual work.     «

This thesis explores the application of Knowledge Graphs (KGs) and Large Language Models (LLMs) to enhance project knowledge management in the construction industry. The research addresses long-standing challenges in knowledge retrieval, organization, and analysis within the sector by proposing a novel framework that leverages the semantic understanding capabilities of LLMs to construct KGs from project communication data. The methodology was tested using an open-source BCF file dataset, represe...     »