High Impact Factor : 4.396 icon | Submit Manuscript Online icon | UGC Approved icon

Threshold Based Association Rule Mining Algorithm for Dynamic Content


Prof. Jyoti Golakia , Atharva College of Engineering, Mumbai University,Mumbai, India; Prof. Trupti Shah, Atharva College of Engineering, Mumbai University,Mumbai, India; Prof. Krishanjali Shinde, Atharva College of Engineering, Mumbai University,Mumbai, India


Data Mining, KDD, ARM


Today, data mining has become a very vast area of research. A data mining technique, Association Rule Mining (ARM), is also a vast area of research. Association rules identify relationships among data items and were introduced in 1993 by Agarwal et al. With the increasing use record-based databases whose data is being continuously added, recentimportant applications have called for the needof incremental mining. In dynamic transactiondatabases, new transactions are appended andobsolete transactions are discarded as timeadvances. Several research works havedeveloped feasible algorithms for derivingprecise association rules efficiently andeffectively in such dynamic databases. Also,sometimes the itemsets are not as frequent asdefined by the threshold, but the associationrules generated from them are still important.Such items are called rare items. Classical ARMframework assumes that all items have the samesignificance or importance which is not alwaysthe case. In this paper, an algorithm – EnhancedNew Fast Update, is proposed for miningassociation rules from dynamic databaseconsidering rare interesting items also.

Other Details

Paper ID: NCTAAP088
Published in: Conference 4 : NCTAA 2016
Publication Date: 29/01/2016
Page(s): 369-373

Article Preview

Download Article