[AIS93] introduced the application of association rules to market basket data. Different data representations such as attribute value pairs and relational data have since been translated into items as originally used in the market basket paradigm.
Apriori Sets And Sequences extends the Apriori algorithm introduces in [AS94]. Specifically the implementation builds upon Apriori Sets [Sho01] which was incorporated into a machine learning and data mining system [FW00] by [SS02].
[Sho01] addresses the idea of having set-valued attributes in addition to single valued attributes from which to mine association rules. A system which translates set valued attributes into items was implemented.
There has been work done that incorporates temporal information into association rules. [AR00] uses the time of a transaction to better identify interesting rules with low support. It also uses this information to declare rules obsolete if the transactions that support a rule are from a certain distance in the past. A similar notion is used in [CP99] to identify periods of time during which a particular association rule is valid. the rules themselves contain no temporal information. [MTV95] explains a method for finding episodes, a collection of events occurring inside the same window of time. Here events refer to instantaneous occurrences occupying a single unit of time. An interesting approach to using temporal information can be found in [FLYH99]. Rules that associate items between transactions where transactions represent successive unit of time are mined. [Dun03] Offers an excellent summary of temporal mining terms and algorithms currently used and known.