Study On Frequent Subtree Sequence Mining

With the development of frequent item and frequent sequence mining, the technology of data mining is gradually to extend its way to solve the problem of structural pattern mining—frequent subgraph(subtree) mining, especially mining the whole frequent subgraph sequence from labelled ordered graph. However, almost no method has been proposed to mine frequent subsequence from a set of tree sequences. At present, with the fast development of the Internet and database, mining of frequent subgraph(subtree) sequence has become meaningful in many applications such as chemistry information, web log, XML, analysis of compounds'structure. In this paper, based on the research of the problem concerning mining frequnet subtree and frequent sequence, and by the definition of the sequence that are composed of subtrees, we tentatively introduce one method that can be used to mine frequent transformation subtree sequence.Firstly, given a sequence that is composed of subtrees, we bring forward a transformation rule between the successive subtrees in the sequence, and then the subtree sequence is presented as its transformation sequence. Otherwise, we also propose an admissibility theorem to ensure the candidate transformation sequence that are produced only.Secondly, the pseudocode of the transformation algorithm is designed, which can be used to transform a subtree sequence to its transformation sequence.In the end, by the definition of the Union Tree and the Proj for a transformation sequence, a mining method called TreeFTSSMiner(Tree Frequent Transformation Sequences Miner) is proposed to mine the frequent transformation sequence and the algorithm is also presented. Furthermore, the result of the experiment is analyzed at the end of the paper. We also bring forward the relative work for the next step.