Research On Mining Algorithm Of Complex Network

Study of complex networks has long been a hotspot. In the field of bioinformatics, we found one after another very meaningful conclusions. But it is not widely used in the field of bioinformatics, it is also used in the fields related to network analysis sociology, computer networks, e-commerce and so on in. The famous "beer and diapers" story, also on commodity sales network derived by mining association rules.The network motif is one of the important of complex networks information mining. Because it can reflect the scale and effectiveness of community groups.Faced with dynamic network network motif mining this problem, first with abstract, formal mathematical language to express the problem. Secondly, analysis of network motif network of conventional dynamic mining algorithm to study the efficiency and derive reason for the low efficiency of the algorithm. And then put forward a new network motifs efficient dynamic network mining algorithm. And proposes three theoretical nature to ensure that the proposed dynamic network motif mining algorithm correctness.Again, according to the nature of the proposed three, designed to achieve a dynamic network motif Mining Algorithm. In algorithm design, re-recorded topology information used to describe the right time to meet the conditions of the sub-network motifs graph, vertex information network and meet the conditions of sub-phantom figure, in order to ensure the integrity of network motifs mining. And theoretical analysis time the new algorithm and the traditional algorithm proposed complexity of the new algorithm than the traditional algorithm efficiency is much higher.Thirdly, through the network for dynamic and static network comparative analysis concluded that the static network is a dynamic network of special circumstances. So that the network molded body dynamic and static network to establish a unified network of mining relations. Also analyzed results, just to the proposed dynamic network motif mining algorithms make the appropriate changes can be applied in the network molded body static network of mining.Finally, through a variety of experiments to verify the previous conclusion that the efficiency of the new algorithm is higher than the efficiency of traditional algorithms. We have designed four experiments, namely:distribution network obey uniform distribution, distribution network obey Poisson distribution, distribution networks and obey Distribution Network Power-Law bipartite graph.