An approach to mining bundled commodities

作者：Yin Yunfei^*

来源：Knowledge-Based Systems, 2008, 21(4): 321-331.

DOI：10.1016/j.knosys.2007.07.003

• bundled commodity
• interval-valued rule
• interval-valued database
• XML structure

Although bundled commodities are largely existed in supermarket, there are few investigations about bundled commodities mining. In this study, interval-valued association rules as a novel and possible approach to solving the bundled commodities mining issue, is proposed. Properties research based on interval-valued association rules is conducted, and an interval-valued rule pattern: < F, boolean OR, boolean AND > is constructed, where F denotes set of interval-valued ruses, boolean OR and boolean AND denote the disjunctive operation and conjunctive operation, respectively. Furthermore, one of the properties satisfied by the interval-valued rules: A boolean AND C = B boolean AND C and A boolean OR C = B boolean OR C <-> A = B, where A, B, C are there different rules, is validated and utilized to mine the bundled commodities. Finally, a large-scale software engineering project relative to interval-valued rule mining is implemented to merge flight testing information about aircrafts which validates the technique of mining bundled commodities can be discovered a special relation between objects. These initial investigations provide a researchable framework for bundled commodities mining.

• 出版日期2008-5
• 单位北京大学