A New Design Knowledge Retrieval Model Based on Granularity and Clustering Theories
Hongwei Liu,Yan Li,Jian Chen,Yan Xiong,Xiaojiao Du,Qian Shi
DOI: https://doi.org/10.1177/0954406216668206
2017-01-01
Abstract:Former research has proved that the design knowledge involved in the conceptual design stage has multiple attributes/perspectives, which can be further abstracted into the concepts with different abstract/granularity levels, from coarse to fine. Based on this, we believe that it is necessary to use multiple attributes to describe design knowledge during the knowledge retrieval process. Firstly, an attribute ontology with multi-perspective and multi-granularity is built in the new retrieval model. Therefore, the knowledge documents can be abstracted by the concepts of the ontology. Based on the correlation of the concepts, the clustering theory is introduced into the new model to cluster knowledge documents, and the combined keyword concepts of the clusters are generated as well. The keyword concepts of the clusters, rather than the keyword concepts of documents, are used as the index of the retrieval. During the knowledge matching process, the keywords and their semantic extension of design problems are extracted, and the similarity between the abstract descriptions of the knowledge clusters and design problems is calculated, so the best cluster can be found by the calculated results. Based on the granularity levels of the keywords (concepts) the design problems, finally the documents in the selected cluster are ranked in the order of granularity levels. The selected document by the model is that with the highest relevance and most suitable granularity level about the design problem. In the last section of the paper, we used a simple real case to evaluate the new model, and also develop the scheme of the whole design knowledge retrieval system for the future work.