Performance Comparison of Hashing Algorithm with Apriori

Several algorithms have already been developed for association rule mining. In Apriori algorithm, if the number of candidate sets are increased, the efficiency of the algorithm decreases. To overcome this, MPIP algorithm, proposes perfect hash function in the initial stages of the algorithm. Here, we propose perfect hash functions for 2- itemsets and 3-itemsets. The function depends on the number of large 1-itemsets. Using hash functions, we propose a hash based method for finding large 2- itemsets and 3-itemsets. This method is compared with apriori method using hash tree and without using hash tree and results are shown.