Mining Association Rules from a Collection of XML Documents using Cross Filtering Algorithm

Since numerous data have been represented and exchanged by XML, the ability to extract useful knowledge from XML data is needed. There are several attempts to mine association rules from XML data. However, they mostly rely on legacy relational database with an XML interface so that efficiency and simplicity are challenging issue. In this paper, HILoP (hierarchical layered structure of PairSet) is introduced. The use of this data structure prevent from multiple XML data scans to mine association rules from a collection of XML documents. Also, cross filtering algorithm is introduced to mine frequent patterns, the algorithm reduces the number of candidate set. The performance evaluation result shows that this mechanism is powerful enough to represent both simple and complex structured association relationships inherent in XML data