Release Time:2019-03-11  Hits:

Indexed by: Journal Article

Date of Publication: 2012-11-01

Journal: International Journal of Advancements in Computing Technology

Included Journals: Scopus、EI

Volume: 4

Issue: 20

Page Number: 625-632

ISSN: 20058039

Abstract: Frequent pattern mining on uncertain dataset takes into consideration of item's existing probability, but does discriminate items by their different importance. To address this issue, we propose a new mining model called High Expected Weight Itemsets Mining (or weighted frequent pattern mining) on uncertain dataset based on the concept of expected weight. We also propose a corresponding algorithm called HEWI-Mine to mine high expected weight itemsets from uncertain dataset using a pattern- growth approach. We perform some testing mining on multiple datasets, and observe different mining results produced. Because items in real-world transaction database do contain different weight properties, such as price or profit, this model may contribute to more precise analysis on business applications involving frequent pattern mining.