A new metric for evaluating the importance of nodes in complex product structure networks

The importance of nodes is a significant metric for the structure of complex product networks. The existing metrics still have deficiencies such as relatively partial consideration of factors, subjective weight distribution among relevant factors, and lack of effective validation schemes. This paper proposed a node importance evaluation method based on Principal Component Analysis (PCA). The local, global, and location attributes of the product network are taken into account. Meanwhile, a node importance verification method based on Spearman’s correlation coefficient was designed by considering the correlation between node importance and its propagation strength. Finally, the proposed algorithm’s effectiveness was verified on ten artificial networks and one real product structure using the SI propagation model.