Mining frequent weighted utility itemsets in hierarchical quantitative databases

摘要

Mining frequent itemsets in traditional databases and quantitative databases (QDBs) has drawn many researchers’ interest. Although many studies have been conducted on this topic, a major limitation of these studies is that they ignore the relationships between items. However, in real-life datasets, items are often related to each other through a generalization/specialization relationship. To consider the relationships and discover a more generalized form of patterns, this study proposes a new concept of mining frequent weighted utility itemsets in hierarchical quantitative databases (HQDBs). In this kind of databases, items are organized in a hierarchy. Using the extended dynamic bit vector structure with large integer elements, two efficient algorithms named MINE_FWUIS and FAST_MINE_FWUIS are developed. The empirical evaluations in terms of processing time between MINE_FWUIS and FAST_MINE_FWUIS are conducted. The experimental results indicate that FAST_MINE_FWUIS is recommended for mining frequent weighted utility itemsets in hierarchical QDBs.