Analysis of Complexities for finding efficient Association Rule Mining Algorithms

Several algorithms for association rule mining, have been implemented including a variation of Apriori, an algorithm using hash functions for finding large 2-itemsets and 3-itemsets and direct search method for finding other large k-itemsets, and another variation of Eclat algorithm using perfect hash functions for 2-itemsets and 3-itemsets and the method of vertical mining for finding other large k-itemsets. All these algorithms were compared, among themselves by finding time complexities of each algorithm, for finding factors affecting the efficiency of the algorithms. By observing the complexities we are able to (i) find factors that is influencing execution times (ii) find why one algorithm works faster than the other and (iii) find convincing method of choosing best algorithm with the given constraints such as database composition, maximum frequent set size and itemset size factors. Keyword: Clustering, classification, and association rules, Data mining, Analysis of algorithms, Graph theory.