Visualization of association rules over relational DBMSs

The focus of this paper is the association rule visualization system that we have designed and developed. The rules produced by the mining algorithm are assumed to be stored in tables. The alternatives for visualization include tabular form, interactive two-dimensional, and three-dimensional graphics. By providing sorting and filtering abilities, the rule visualization system proposed in this paper provides a flexible, efficient, and easier way to manage and understand large number of association rules. As a result, this visualization system becomes an essential part of our association rule mining subsystem. We compare our association rule software with Intelligent Miner from IBM in various aspects, such as data accessibility, user interface, input/output, and rule visualization.