[r] Bayardo98, "Efficient Mining Long Patterns from Databases"

@inproceedings{276313,
author = {Roberto J. Bayardo, Jr.},
title = {Efficiently mining long patterns from databases},
booktitle = {SIGMOD '98: Proceedings of the 1998 ACM SIGMOD international conference on Management of data},
year = {1998},
isbn = {0-89791-995-5},
pages = {85--93},
location = {Seattle, Washington, United States},
doi = {http://doi.acm.org/10.1145/276304.276313},
publisher = {ACM Press},
address = {New York, NY, USA},
}

Motivation:
To address aprior's limitations in mining long patterns.
Contribution:
Proposed an effective approach to find Maximal Frequent itemsets without loss of completeness. Performance is one to two orders of magnitude better than apiori-like algorithms.
Methods:
Use item iteration tree to present all itemsets.
Do superset and subset pruning.
Use lower-bound support to save database scanning by using subset support information.
Discussions:
This approach can find maximal frequent itemsets efficiently. However, all frequent itemsets are implicit (though completely contained) in the caculated results. For tasks like association rule mining, there will still be substantial computation.