Higher Order Mining: Modelling And Mining TheResults Of Knowledge Discovery

To date, most data mining algorithms and frameworks have concentrated on the extraction of interesting rules directly from collected data. This paper investigates the generic modelling of these rules and the utility of deriving rules from the results of other data mining routines, that is, mining from rulesets (or meta-mining). It is argued that this approach has three signi cant advantages. Firstly, with the expansion of dataset size, the tractability of mining from the complete dataset may be diAEcult on a regular basis, secondly, changes in observations (and therefore in the observed system) can be more easily discovered by inspecting changes in extracted rules over time (or over any other sequential progression), and nally, the nature of the rules extracted by this process are that they contain di erent higher order semantics from that exhibited by rst order discovery process. We argue that, in many cases, such rules are closer to the sorts of rule frequently used to describe everyday phenomena.