摘要
关联规则获取是知识发现和数据挖掘中的核心问题之一。对超市来讲,从交易数据中挖掘出的关联规则有两点重要意义:一是有助于设计商品的摆放位置;二是帮助商品进货搭配规划,为更好利用关联规则进行进货搭配规划,知识工程师不仅需要考虑关联规则的可信度、支持度和兴趣度,更需要考虑支持集对关联规则的贡献度和关联规则自身的平衡度和复杂度。本文首先采用形式概念分析理论挖掘交易数据中的关联规则,这些规则具有100%的可信度。然后,在关联规则柔性筛选的基础上进行商品进货决策规划。所谓柔性是指用户可自己定义规则的不同阈值组合(例如析取和合取)选择规则。
One of the core tasks of knowledge discovery and data mining is the mining of association rules. For some supermarket, they mainly focus on the following two important significances of association rules acquired from a lot of trade-off data： one is that they can be used to arrange the places of correlative products on the shelves, and the other is that they can be used to plan harmonious stocks of the correlative products. In order to settle the aims, knowledge engineers not only should take into account confidences, supports and interest degree of association rules but focus on contribution degree of support set and balance/complexity of association rules. In the paper, the theory of Formal Concept Analysis are firstly used to acquire association rules with confidence 100% from a large of trade-off formal context, and then discussed the products planning based on flexible filtering association rules. Flexibility means that users can define different thresholds combination of rules （e. g. disjunction and conjunction） for choosing satisfactory rules.
出处
《计算机科学》
CSCD
北大核心
2008年第1期213-215,235,共4页
Computer Science
基金
山东省优秀中青年科学家奖励基金“图的因子理论及其在网络技术中的应用”(编号2005BS01016)资助
关键词
形式概念分析
关联规则
柔性筛选方法
Formal concept analysis
Association rule
Flexible filtering method
