ABSTRACT
Knowledge of design rationale provides support for design reuse, innovative design, collaborative design and other design behaviors. It is difficult to guarantee the accuracy and universality of design knowledge in a single instance of design rationale, and it is difficult for designers to reuse the design knowledge directly. This paper proposes a design rationale knowledge-discovery method based on frequent-pattern mining. In order to facilitate the mining of frequent patterns from a fine-grained design rationale model, we first transform it into a label graph, then extract the frequent patterns from the label graph, and then transform these into a design rationale model. Then, through the design rationale model fragment integrity detection and inference method, the extracted design rationale model structure is perfected as a complete design rationale model fragment. Finally, the validity of this method is illustrated by way of an example of design rationale models for an automatic marking machine.
