Discovering Higher Level Correlations from XML Data

This chapter proposes the XML-GERMI framework to support XML data analysis by automatically extracting generalized association rules (i.e., higher level correlations) from XML data. The proposed approach, which extends the concept of multiple-level association rules, is focused on extracting generalized rules from XML data. To drive the generalization phase of the extraction process, a taxonomy is exploited to aggregate features at different granularity levels. Experiments performed on both real and synthetic datasets show the adaptability and the effectiveness of the proposed framework in discovering higher level correlations from XML data.