关键词: 数据挖掘, 动态关联规则, 趋势度, 元关联规则

Abstract: Based on the original definition and classification of Support Vector (SV) and confidence vector, this paper put forward a method of data tendency measure mining in dynamic association rules, according to the characteristic of rules with time changing. First, taking advantage of tendency measure threshold to eliminate useless rules, the item sets candidates can be reduced. Second, producing the dynamic association rule, this method found out valuable rules and improved the mining quality. Finally, by analyzing a transaction database that is characterized by the tendency of changes and cycles, the analytical results verify the validity of the proposed method.

Key words: data mining, dynamic association rule, tendency measure, meta-association rule