A New Cosine Similarity Matching Model for Interior Design Drawing Case Reasoning

Abstract

In our previous study, we proposed a recommender system for interior design drawing retrieval. In that paper, a cosine similarity matching function is used for measuring binary bit string similarity between two design cases. After wider practical applications, we found that the design features on the interior design drawings could be with mixed interval, nominal, ordinal or ratio measurement scales. We further found that a case-based reasoning system is more suitable for interior design drawing retrieval than a recommender system, because the case-based reasoning system can begin with a few number of reference cases and allows the case database to be developed incrementally. Therefore, the objective of this study is to propose a new cosine similarity matching model for interior design drawing case retrieval in a case-based reasoning system, in which mixed measurement scales are considered and applied. Finally, a numerical case study is carried out to demonstrate the effectiveness and capabilities of the proposed cosine similarity matching model for the case-based reasoning system.