Bim Oriented Intelligent Data Mining and Representation

Building Information Model (BIM) encompasses mass of data specified by Industry Foundation Classes (IFC) from multiple fields during the whole lifecycle of the Architecture, Engineering and Construction (AEC) project. Recently, a large international initiative is launched to provide extensive support that will facilitate the creation, sharing and integration for BIM through all professions in the AEC industry, while few bend their mind to intelligent BIM data mining, which will improve the value of BIM. As the Information becoming increasingly rich and complicated, utilization of the data is getting harder, particularly for personnel without extensive knowledge of IFC specifications. This paper proposes a framework utilizing Natural Language Processing (NLP) and International Framework for Dictionaries (IFD) to address the intelligent BIM data mining problem. First, user requirements depicted with a sentence of natural language are processed for keywords extraction by NLP. Then, keywords will map to IFC entities (or properties) through IFD-driven mapping mechanism, providing information for data retrieval and analyzing. Finally, the analyzed results of BIM data collected by previous research and application will be represented in accordance with its format such as tables, charts, animations or combination of them. The framework provides an intelligent data mining and multi-aspect data representation method for users without any special training, thus enhancing the applicable value of BIM. Practical application results in construction management illustrates that with semantic understanding of his/her intention in natural language, user concerned data will be automatically retrieved, analyzed and represented in a suitable form, which is of great benefit for corporations without requiring extremely technological users, facilitating BIM application and enhancing the value of BIM.