k-PFPMiner: Top-k Periodic Frequent Patterns in Big Temporal Databases

Finding periodic-frequent patterns in temporal databases is a prominent data mining problemwith bountiful applications. It involves discovering all patterns in a database that satisfy the user-specifiedminimum support(min_sup) and maximum periodicity(max_per) constraints.Min_supcontrols the leastnumber of transactions in which a pattern must appear in a database.Max_percontrols the maximumtime interval within which a pattern must reappear in the database. The popular adoption of this task hasbeen hindered by an open problem, which involves setting appropriatemin_supandmax_pervalues forany given database. This paper addresses this open problem by proposing a solution to discover top-kperiodic-frequent patterns in a temporal database. Top-kperiodic-frequent patterns represent theknumberof periodic-frequent patterns having the lowestperiodicityvalue in a database. An efficient depth-first searchalgorithm, Top-kPeriodic-Frequent Pattern Miner (k-PFPMiner), which takes onlykthreshold as an input,was presented to find all desired patterns in a database. Experimental results on synthetic and real-worlddatabases demonstrate that our algorithm is efficient and scalable.

Automated summary

Finding periodic-frequent patterns in temporal databases is a significant data mining problem. This paper proposes a solution to discover the top-k periodic-frequent patterns in a database.