Research And Application Of Frequent Subgraph Mining Algorithm

Data mining technology has been a hot task in the field of database and artificial intelligence in recent years, and it has attracted extensive attention in science and technology industry. Abundant literature has been dedicated to this research and tremendous progress has been made.With the deeply study of the frequent pattern mining, graphs can modeling for many transactions widely, and the study of graphs have become increasingly important. The dissertation first introduces the definations and background of subgraph mining. Then it introduces and analyses the states of domestic and foreign researchs. For the study, it gives a detailed introduction of the basic knowledge and technical.Based on FP-Tree, This dissertation presents an improved FP-growth algorithm, which called CFPG, can find the closed frequent connected subgraph from the model of unique labeled directed connected graph set. The experiment of biology metabolize pathway dataset demonstrated that the algorithm can get the closed frequent subgraph set effectively, and can get the max frequent subgraph sets of many different threshold by execute once. This algorithm can use for mining the network or graph set which can modeling by unique labeled, directed, connected graph.For the labeled undirected graph, the dissertation presents a frequent subgraph mining algorithm ,which called FSDE, based on Apriori and technology of embeddingset. The algorithm also improves the canonical label and candidate generation. The experiment of PTE dataset demonstrated that the algorithm can get all frequent subgraph sets effectively, and more effective than the Apriori-based algorithm FSG.