Pattern Discovery in Temporal Databases: A Temporal Logic Approach

1 This work was supported in part by the NSF under grant IRI-93-18773. Abstract The work of Mannila et al. [4] of finding frequent episodes in sequences is extended to finding temporal logic patterns in temporal databases. It is argued that temporal logic provides an appropriate formalism for expressing temporal patterns defined over categorical data. It is also proposed to use Temporal Logic Programming as a mechanism for the discovery of frequent patterns expressible in temporal logic. It is explained in the paper how frequent temporal patterns can be discovered by constructing temporal logic programs. To test these methods , temporal logic programs were constructed for certain classes of patterns and were implemented in OPS5.