An Efficient Algorithm to Mine Unordered Trees

Mining unordered trees are very useful in domains like XML date, biological information, Web structure, etc. In this paper, we introduce an efficient algorithm UTMiner (unordered trees miner). As the trees are unordered, in order to avoid mining the same subtrees, an efficient unordered trees standardization is first introduced to transform the unordered trees into the standard subtrees. Then UTMiner is used to get all standardized subtrees. UTMiner builds a multilayered data structure based on subtree vector and the hash table so it reduces isomorphism time in the mining process. It requires only one database scanning so it reduces the scanning times and improves the efficiency, particularly in a large databasepsilas mining process. Many experiments have shown that the UTMiner is feasible and more efficient than other.