A Novel Algorithm for Temporal Infrequent Weighted Itemset Mining using FP Growth

The Infrequent Weighted Itemet system focuses on the issue of discovering infrequent itemsets by using weights for differentiating between relevant items and not within each transaction. To reduce the complexity of the mining process in high dimensional databases, the temporal infrequent weighted itemset TIWI Miner adopts an FP-tree node pruning strategy to early discard items (nodes) that could never belong to any itemset satisfying the TIWI-support threshold. The SMA, a split and merge algorithm for infrequent item set mining, which can easily be extended to allow for “fast data� mining in the sense that dynamic data. Other distinguishing qualities of the method are its exceptionally simple processing scheme and data structure, it very easy to implement, convenient to execute on dynamic and external storage. The algorithm integrates a novel strategy named EUCI (Estimated Utility Co-occurrence Identification) to reduce the number of joins operations when mining low-utility item sets using the SM (split and Merge) data structure.