Efficient frequent subgraph mining algorithm

With the successful development of frequent item set and frequent sequence mining, the technology of data mining is natural to extend its way to solve the problem of structural pattern mining-Frequent subgraph mining. Frequent patterns are meaningful in many applications such as chemistry, biology, computer networks, and World-Wide Web. This paper proposes a new algorithm GraphGen for mining frequent subgraphs. GraphGen reduces the mining complexity through the extension of frequent subtree. For the best algorithm available, the complexity is O(n³·2ⁿ), n is the number of frequent edges in a graph dataset. The complexity of GraphGen is O(2n·(n^2.5/logn)), which is improved O(√n·logn) times than the best one. Experimental results prove this theoretical analysis.