Towards Efficient Discovery of Periodic-Frequent Patterns in Dense Temporal Databases Using Complements

Periodic-frequent pattern mining involves finding all periodically occurring patterns in a temporal database. Most previous studies found these patterns by storing the temporal occurrence information of an item in a list structure. Unfortunately, this approach makes pattern mining computationally expensive on dense databases due to increased list sizes. With this motivation, this paper explores the concept of complements, and proposes an efficient algorithm that records non-occurrence information of an item to find all desired patterns in a dense database. Experimental results demonstrate that our algorithm is efficient.