Publication

Abstract : One of the most common problems in data mining is to find frequent itemsets. There are various algorithms which extract such itemsets from large database based on Minimum Support Threshold (MST). They further generate association rules based on Minimum Confidence Threshold (MCT). These two threshold values are defined by user or organization. Apriori is one of the most popular data mining algorithms but it generates all frequent itemsets and association rules which may be of user's interest or may not be. Proposed algorithm prunes all uninteresting frequent itemsets generated in every level and only considers those items and rules which are of interest based on MST and MCT. It saves considerable storage space and time. Every level prunes such items and rules which are of no interest and forwards the resultant list to next iteration.