Discovering Periodic-Frequent Patterns in Uncertain Temporal Databases

R. Uday Kiran, Palla Likhitha, Minh-Son Dao, Koji Zettsu, Ji Zhang

Abstract: Periodic-frequent pattern mining aims to discover all periodically occurring frequent patterns in a temporal database. Most previous studies focused on finding these patterns by disregarding the items’ uncertainty nature in the data. This paper proposes a novel model of periodic-frequent patterns that exist in an uncertain temporal database. We introduce a new tree-structure and an algorithm to find all desired patterns in the database effectively. We have also presented two tighter upper bound measures to reduce the computational cost effectively. Experimental results on various databases demonstrate that our algorithm is memory and runtime efficient.

External IDs:dblp:conf/iconip/KiranLDZZ21