Abstract: The concept lattice is an effective tool for knowledge discovering and knowledge processing, which has been successfully applied in many fields. However, lattice constructions with reduction operation are mostly based on minimum support and degree of confidence. At the same time, the degree of confidence is used to extract rules. This paper proposes the Hasse diagram with strong association which is constructed by the mutual information of information theory, and extracts rules by mutual information.

Key words: strong association rules, concept lattice, mutual information, rule extraction, data mining

摘要： 概念格是一种有效的知识表示和知识发现的工具，已被成功应用于许多领域，然而在建格上大多是利用最小支持度以及置信度来进行约简操作，同时利用置信度来进行规则提取。提出以信息论的互信息来构造具有强关联规则的Hasse图，并利用互信息进行规则提取。

关键词: 强关联规则, 念格, 互信息, 规则提取, 数据挖据