基于频繁子图挖掘的典型零件结构获取方法

A Typical Part Structure Acquisition Method based on Frequent Subgraph Mining

典型零件结构是产品的隐性设计知识,用于企业零件资源聚类和检索。为了从零件库中获取典型零件结构,提出基于频繁子图挖掘的典型零件结构获取方法。从基于特征的三维零件模型中提取和筛选特征关系,并建立其有向特征关系图;由海量三维零件模型的有向特征关系图构成有向特征关系图库;利用Apriori频繁子图挖掘算法对有向特征关系图库实施频繁子图挖掘;通过频繁子图与零件库之间映射关系获取典型零件结构。实例验证了该方法的可行性。

Typical part structure is a kind of tacit knowledge in product design,and was applied to parts clustering part retrieval and etc.In order to acquire typical part structure from similar parts,a novel method based on frequent subgraph mining was proposed Firstly feature relations are abstracted and selected from feature-based 3D part model,and directed feature relationship graph（DFRG） was constructed according to the feature relations,and massive DFRGs comprise the DFRG base.Secondly,an Apriori-based algorithm was designed to mine frequent subgraphs from the DFRG base Thirdly the part structures corresponding to subsequent subgraphs mined were acquired from 3D part base and all these part structures were typical part structure.Finally the feasibility of proposed method was verified by an application example.