Description

Mining frequent itemsets is an active area in data mining that aims at searching interesting relationships between items in databases. It can be used to address to a wide variety of problems such as discovering association rules, sequential patterns, correlations and much more. A transactional database is a data set of transactions, each composed of a set of items, called an itemset (frequently occuring in a database ). Existing methods often generate a huge set of potential high utility item sets and their mining performance is degraded consequently. There is a lacking of mining performance with these huge number of potential high utility itemsets ; higher processing Time too. Two novel algorithms as well as a compact data structure for efficiently discovering high utility itemsets are proposed. High utility itemsets is maintained in a Catalog-based data structure named UP- (Utility Pattern Catalog). Implementing mining process through Discarding Local Unpromising Items and Decreasing Local Node Utilities strategies. Finally Heuristic rule framing is done with respect to the datasets. On adopting this rule framing strategies the strength of the item sets are evaluated. Thus our proposed mechanism scales for any number of transactional log sets. Experimental results predicts that not only reduce the number of candidates effectively but also outperform other algorithms