Build2Vec: Building Representation in Vector Space
This work addresses the problem of analyzing building data for applications in architecture and construction, though it appears incremental as it applies an existing method (node2Vec) to a new domain (BIM data).
The paper tackles the problem of representing building components from Building Information Models (BIM) by developing Build2Vec, a graph embeddings algorithm that uses node2Vec with biased random walks to transform labeled property graphs into multi-dimensional vectors. The result is a promising approach demonstrated in a case study on a net-zero-energy building at the National University of Singapore, showing it can capture semantic relations and similarities of building objects, particularly spatial and spatio-temporal data.
In this paper, we represent a methodology of a graph embeddings algorithm that is used to transform labeled property graphs obtained from a Building Information Model (BIM). Industrial Foundation Classes (IFC) is a standard schema for BIM, which is utilized to convert the building data into a graph representation. We used node2Vec with biased random walks to extract semantic similarities between different building components and represent them in a multi-dimensional vector space. A case study implementation is conducted on a net-zero-energy building located at the National University of Singapore (SDE4). This approach shows promising machine learning applications in capturing the semantic relations and similarities of different building objects, more specifically, spatial and spatio-temporal data.