Visualizing Frequent Itemsets, Association Rules, and Sequential Patterns in Parallel Coordinates

摘要

Frequent itemsets, association rules and sequential patterns are defined on elements of power sets of items and reflect the many-to-many relationships among items. Although many tools have been developed to visualize association rules, none of them can simultaneously manage a large number of rules with multiple antecedents and multiple consequences. This problem is shown as a straightforward application of parallel coordinates. We show that, by properly arranging items on coordinates and by filtering out subsets of large frequent itemsets, item groups can be naturally displayed and that inherent properties such as partial orders in itemsets and in association rules are implied by this visualization paradigm. The usefulness of this approach is demonstrated through examples.