Simple yet efficient approach for maximal frequent subtrees extraction from a collection of XML documents

Recently, XML is penetrating virtually all areas of computer science and information technology, and is bringing about an unprecedented level of data exchange among heterogeneous data storage systems. With the continuous growth of online information stored, presented and exchanged using XML, the discovery of useful information from a collection of XML documents is currently one of the main research areas occupying the data mining community. The mostly used approach to this task is to extract frequently occurring subtree patterns in trees. However, the number of frequent subtrees usually grows exponentially with the size of trees, and therefore, mining all frequent subtrees becomes infeasible for a large tree size. A more practical and scalable approach is to use maximal frequent subtrees, the number of which is much smaller than that of frequent subtrees. Handling the maximal frequent subtrees is an interesting challenge, and represents the core of this paper. We present a novel, conceptually simple, yet effective approach that discovers maximal frequent subtrees without generation of candidate subtrees from a database of XML trees. The beneficial effect of our approach is that it not only reduces significantly the number of rounds for infrequent tree pruning, but also eliminates totally each round for candidate generation by avoiding time consuming tree join operations or tree enumerations.