Data mining and recommendation of engineering note items in MBD dataset

To meet the requirement of product full 3D digitalization development, a data mining and recommendation method based on the association rules about engineering note items in MBD (Model Based Definition) dataset is proposed. The helpful knowledge and experience can be obtained from MBD dataset's creation history based on association rule theory in data mining. In this method, all the design, manufacturing and inspection standards and information used in the product’s development process, which can be called engineering note items, are analyzed, decomposed, encoded, managed and released. Then FP-growth algorithm is taken to get association rules from the MBD dataset's history records, and the potential association relationship among engineering note items can be revealed. Finally, the recommendation and the completeness check about the engineering note items are realized based on the association rules. The MBD dataset definition system has been developed and implemented in an enterprise based on this method. The practice is proved that this method could effectively reveal the latent association rules from the history records and the MBD dataset's creation efficiency and quality by engineering note items recommendation is improved.