Demonstration: A Prototype for the Exploratory Mining of Constrained Association Rules

Welcome to a demonstration of our CAP prototype. The CAP prototype demonstrates a new exploratory model of association mining where the mining process is split up into two phases. In Phase I, users specify constraints on the data to be mined. In Phase II, users can manually select only those candidate pairs that are interesting enough to find relationships on, and proceed to select from three different relationship metrics to assess candidate association.

More detailed information regarding the exploratory model, the CAP algorithm, and this demonstration can be found in the handout given at SIGMOD 99. The following lists the demo screens in the order that they should be viewed. Note: This demonstration is best viewed with a resolution of 1024 x 768.

Introduction to the Demonstration

1. CAP - Getting Started
2. Opening a Database File
3. Viewing the Metadata
4. Viewing the Relation
5. The Main Window
6. Creating Antecedent Constraints
7. Antecedent Class Constraint
8. Antecedent Frequency Constraint
9. Antecedent Domain Constraint
10. Creating Consequent Constraints
11. Consequent Class Constraint
12. Consequent Frequency Constraint
13. Consequent Aggregate Constraint
14. Two-Variable Constraints
15. Running CAP
16. One-Variable Frequent Sets
17. Two-Variable Candidate Pairs
18. Sorting Frequent Sets
19. Chart of Sorted Candidate Sets
20. Sorted Frequent Sets
21. Candidate Set Selection Mode
22. The Selection Helper
23. Specifying the Relationship Metric and Significance Threshold
24. Related Sets
25. Reiterating Through Phase II
26. Dependent Sets