Mining association rules at multiple concept levels leads to the discovery of more concrete knowledge. Nevertheless, setting an appropriate minsup for cross-level itemsets is still a non-trivial task. Besides, the mining process is computationally expensive and produces many redundant rules. In this work, we propose an efficient algorithm for mining generalized association rules with multiple minsup. The method scans appropriately k+1 times of the number of original transactions and once of the extended transactions where k is the largest itemset size. Encouraging simulation results were reported.
|Name||Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|
|Conference||4th European Conference on Principles and Practice of Knowledge Discovery in Databases, PKDD 2000|
|Period||13/09/00 → 16/09/00|
- Theoretical Computer Science
- Computer Science(all)