From patterns to evidence: Enhancing sustainable building design with pattern recognition and information retrieval approaches

Abstract

Decision-making in design and engineering relies little on knowledge discovered in previous projects and embedded in digital data. Applying analytical computational techniques to available data and pro- cesses can be of significant influence for infusing decision-making with the evidence-based character that it is currently lacking. The design environment is where decisions are implemented, therefore, we aim to endow it with knowledge discovered in previous projects and existing buildings. We use an approach that combines data mining and semantic modelling for case-based design (CBD). We investigate the character of the active design environment, what queries can be constructed automatically from the data available in that environment, and how they can be executed against a repository of design models and performance patterns obtained using Knowledge Discovery in Databases (KDD) and various machine learning approaches. We demonstrate this approach on a use case, highlighting its potential for evidence-based design decision support.