The objective of data mining is to discover useful and interesting relationships among the data in a database. Temporal information can be used to provide a linear ordering on the occurrence of events, to determine inter-event relevance, and to link events in a data stream. Events may be linked by the source of the information or by shared attributes. Event linking permits the discovery of relationships that occur among groups of events, rather than restricting the relationships to those that occur within a single event and extends the type of relationships that can be represented and analyzed in a temporal event stream. In this paper, we examine modifications to standard data mining algorithms suitable for identifying sporadic relationships in temporal data defined using event linking and fuzzy relevance constraints.