Journal of Theoretical and Applied Information Technology
Transaction database is used to find frequent itemsets from large datasets along with their associated timestamps. The existing frequent pattern mining algorithms such as Apriori, FP-growth, Partition, and Pincer Search do not consider the timestamps associated with the transactions. In real time transactions, without timestamps it is difficult to identify the constantly changing behaviour of the frequent patterns in a transaction database. To overcome the above problem, EP-TP (an Efficient Periodic Transitional Patterns) mining method is introduced in this paper. Transitional patterns are used to discover transi-frequent patterns along with their time stamps in a transaction database.