A COMBINTORIAL TREE BASED FREQUENT PATTERN MINING

Abstract

Frequent pattern mining is a process of extracting frequently occurring itemset patterns from very large data storages. These frequent patterns are used to generate association rules which define the relationship among items. The strength of the relationship can be measured using two different units namely support value and confidence level. Any relationship that satisfies minimum threshold of support value is known as frequent pattern. There are several methods and algorithms suggested to mine frequent patterns from large databases. Most of the methods can be assessed for its complexity based on the number of processing levels and number of candidate sets with subsets that are generated in each level. In this study, the combinatorial approach which generates minimal number of combinations using a tree structure and automatically filters infrequent itemsets and mine frequent patterns is suggested. It scans input database once and carries out minimized intersections to count the support value. The complexity is based on the number of transactions and the maximum length of transactions. The new approach purely depends on the size of input transaction database. The combinatorial approach does not depend on the unknown number of processing levels and there is nocandidate sets and subsets generation. The proposed method makes minimal number of combinations when compared to number of candidate sets and subsets in other methods. The method is compared with number of existing legendary methods for its performance.