Improving the efficiency inDiscovering high utility itemsetsUsing parallel processing methods

Abstract

Frequent Itemset Mining (FIM) is a noteworthy subdivision of datamining. FIM identifies frequent patterns by considering the number ofoccurrences of particular items in the transactions of a transaction database.An itemset is deliberated as frequent only if its support count fulfils theminimum support threshold value that the user decides. The FIM algorithmstudies only the number of occurrences of items in the transaction. Therenowned algorithms for frequent itemsets mining are Apriori, FrequentPattern (FP) growth, ECLAT etc. Later, several variants of FIM have beenaddressed such as periodic frequent itemset mining, infrequent itemsetmining, time series based periodic frequent itemset mining etc. newlinePeriodic Frequent Itemset Mining (PFIM) uncovers itemset thatoccur periodically in the transactions. The term Periodicity means thetendency to occur at regular intervals. The occurrence may be bounded withannual occurrence, seasonal occurrence, transaction intervals based ontransaction number alone without any time bounds etc. The need to associateperiodicity in mining is to find the period of frequent item which may help to newlineenhance the business by increasing the stock of frequent items in thatparticular time period. Several works have been carried out to find periodicpatterns in the transaction database that even borrows idea from the FIMworks existing in the literature.Infrequent Itemset Mining (IIM) is a concept that discovers rarecase of itemsets. In contrast with FIM, infrequent itemsets are those whosefrequency falls below minimum threshold value supplied by the user. Eventhough these itemsets are infrequent in occurrence but has proven its potential newlinein the field of bio-informatics newline newline