Data mining for association rules aims to discover interesting relationships among sets of items in a database. Very often these databases include some kind of temporal information, the most common being a temporal label indicating transaction date. Within the field of association rule mining temporal information has been used to obtain sequential association rules, periodic or cyclic association rules, calendric association rules, or event-driven association rules. The temporal component is also relevant when analyzing how association rules evolve if datasets are evaluated on different time-slices. On the other hand, in traditional association rules item attributes were usually Boolean, but many attributes in current databases are quantitative in nature. Fuzzy temporal association rules arise from the use of fuzzy sets to describe quantitative temporal and/or not temporal attributes of items in a database, and/or to introduce fuzzy temporal specifications for the rules a user is interested in; the use of fuzzy sets allows a linguistic interpretation of the rules and also provides means to handle the uncertainty present in attribute measurements. Depending on the rule pattern the final user is interested in, different methods for fuzzy temporal association rule mining can be found in the literature, with mining algorithms adapted to the rule model being used. (C) 2013 John Wiley & Sons, Ltd.
展开▼
机译：关联规则的数据挖掘旨在发现数据库中项目集之间的有趣关系。这些数据库通常包含某种时间信息，最常见的是指示交易日期的时间标签。在关联规则领域中，挖掘时间信息已用于获取顺序关联规则，周期性或循环关联规则，日历关联规则或事件驱动的关联规则。如果在不同的时间片上评估数据集，则在分析关联规则如何演变时，时间成分也很重要。另一方面，在传统的关联规则中，项目属性通常是布尔值，但是当前数据库中的许多属性本质上都是定量的。模糊时间关联规则源于使用模糊集描述数据库中项目的定量时间和/或非时间属性，和/或为用户感兴趣的规则引入模糊时间规范;使用模糊集可以对规则进行语言解释，还可以提供处理属性度量中存在的不确定性的方法。根据最终用户感兴趣的规则模式，可以在文献中找到用于模糊时间关联规则挖掘的不同方法，并使用适合于规则模型的挖掘算法。 （C）2013 John Wiley＆Sons，Ltd.
展开▼
