On Efficient and Effective Association Rule Mining from XML Data

Abstract

In this paper, we propose a framework, called XAR-Miner, for mining ARs from XML documents efficiently and effectively. In XAR-Miner, raw XML data are first transform ed to either an Indexed Content Tree (IX-tree) or M ulti-relational databases (Multi-DB), depending on the size of XML document and memory constraint of the system, for efficient data selection in the AR mining. Concepts that are relevant to the AR mining task are generalized to produce generalized meta-patterns. A suitable metric is devised for measuring the degree of concept generalization in order to prevent under-generalization or over-generalization. Resultant generalized meta-patterns are used to generate large ARs that meet the support and confidence levels. An efficient AR mining algorithm is also presented based on candidate AR generation in the hierarchy of generalized meta-patterns. The experiments show that XAR-Miner is more efficient in performing a large number of AR mining tasks from XML docume nts than the state-of-the-art method of repetitively scanning through XML documents in order to perform each of the mining tasks.