부산대학교 도서관

As a high incidence of chronic inflammatory skin disease, acne has a complex etiology and pathogenesis, and microbial colonization is currently considered as one of the important causes. Therefore, in this paper metagenomics data is analyzed using Frequent-Pattern Growth (FP-Growth) method for aided diagnosis of acne disease. The main ideas include that firstly, the data sets are transformed into binary form of 0 or 1. The original data of lipids whose component content is 0 is set 0, and the original data of lipids whose component content is not 0 is set 1. Then the data sets are scanned to build a frequent pattern tree (FP-tree) based on the frequency of each element and support. Finally, FP-tree is used to determine frequent itemsets. The element items in frequent itemsets correspond to the lipids which are highly correlated with the corresponding data. Experimental results on dataset including normal control (NC), acne healthy skin (HS) and acne diseased skin (DS) show that the proposed method can determine the frequent itemsets of different sample sets. Lipids that can distinguish different skin states are also determined by comparing the difference of frequent itemsets, which can provide guiding help for the auxiliary analysis and treatment of skin acne.