An integrated system of text mining technique and case-based reasoning (TM-CBR) for supporting green building design

Green building has been commonly accepted as an important strategy adopted by governments around the world for mitigating climate change and energy shortage problem. However, the selection and application of green building technologies under different situations usually puzzles designers, although various advanced technologies for green building are available. This study therefore introduces an integrated system of text mining and case-based reasoning (TM-CBR) to help designers retrieve the most similar green building cases for references when producing design for new green buildings. It is the first attempt in this study to integrate text mining technique into a CBR system to improve the efficiency of decision making in green building design. There are two major components of TM-CBR, case representation and case retrieval. Two kinds of case features, namely, identified features and textual features are used collectively to represent a green building case. Four value formats are considered to measure local similarity in the process of case retrieval. Seven cases are chosen randomly from 71 LEED collected cases as the target cases to test the effectiveness of the TM-CBR system. This study provides a new approach to retrieve the successful experience from similar previous cases to improve the effectiveness and adequacy of green building design.