Lost in Translation: Achieving semantic consistency of name-identity in BIM

Abstract

Custom room naming in architectural projects can vary considerably depending on the user. Having multiple and diverse names for the same room is particularly problematic for information retrieval processes in BIM-based projects. Current best practice includes either team agreement on naming labels in BIM or manual renaming to align with an office-wide standard. Both remain laborious and flawed and lead to compounding errors. This research explores how an automated naming-standardization workflow can enhance the interoperability of object-based modeling in a BIM environment and make information retrieval more reliable for a project life cycle. This paper presents research on (1) building a custom corpus specialized for architectural terminology to fit into the BIM environment and (2) devising a standard-naming system titled WuzzyNaming to save manual work for BIM users in maintaining room-name consistency. Our presented workflow applied natural language processing (NLP) technique and Fuzzy logic to perform the semantic analysis and automate the BIM room-name standardization.