Mining actionable combined high utility incremental and associated patterns

Publication Type:
Conference Proceeding
Citation:
AUS 2016 - 2016 IEEE/CSAA International Conference on Aircraft Utility Systems, 2016, pp. 1164 - 1169
Issue Date:
2016-11-17
Full metadata record
Files in This Item:
Filename Description Size
07748234.pdfPublished version330.69 kB
Adobe PDF
© 2016 IEEE. High Utility Itemsets(HUI) Mining, instead of Frequent Pattern Mining (FIM), has been an attractive theme in data mining domain for over a decade since it can be regarded as an alternative way for researchers to identify actionable patterns. In addition, the necessity of decision-making actions and behavior-oriented strategies based on large amount of informative data impels the significance of discovering actionable patterns to be widely admitted. The current HUIM research focus has been on improving the efficiency to make algorithms faster and more stable. However, the coupling relationships between items in given itemsets are ignored. For example, the utility of one itemset might be lower than the manager expected until one additional item takes part in; and vice versa, the utility of an itemset might drop sharply when another one joins in. What's more, it is not occasional to find out that quite a lot of redundant itemsets sharing the same underlying item are presented based on existing academic HUI mining methods. Store managers would not make expected profits based on such results which makes the results not actionable at all. To this end, here we introduce a new framework for mining actionable patterns, called Mining Utility Associated Patterns (MUAP), which aims to find high utility incremental and strongly associated item/itemset with combined incorporating criteria. The outputs of this algorithm are convincing on real datasets as well as synthetic datasets.
Please use this identifier to cite or link to this item: