Algorithm of maximum frequent sub-tree mining based on graph data

In view of the the current frequent subtree mining problems, the maximal frequent subtree mining algorithm can improve the efficiency of the frequent subtree mining algorithm. Firstly, propose two operations of the join and extension on the basis of effective coding, which are generate all candidate subtrees. Secondly, the embedded set is used to solve the problem of subtree isomorphism, and a new maximum frequent subtree mining algorithm(MFST) is proposed. Finally, the correctness of the algorithm is proved and the time performance of the algorithm in the worst case is analyzed. And compared with other frequent subtree mining algorithms based on semi-structured data sets, prove this algorithm is superior to the current frequent subtree mining algorithms. The experimental results show that the MFST algorithm has better time and spatial performance, and can effectively mining frequent subtree on the graph data sets.