Weighted Support Association Rule Mining using Closed Itemset Lattices in Parallel

Summary In this paper, we propose a new algorithm which associates weight to each item in the transaction database based on the significance of the corresponding item. Weighted support is calculated using the weight and the frequency of occurrence of the item in the transactions. This weighted support is used to find the frequent itemsets. We partition the database among ‘N’ processors and generate closed frequent itemsets in parallel. The parallel algorithm used minimizes communication by exchanging only weighted supports among the processors. We generate closed frequent itemsets to reduce the number of itemsets and also as all frequent itemsets can be obtained from closed frequent itemsets, we are not losing any interesting and significant itemsets. The performance of the proposed algorithm is compared to count distribution algorithm in terms of scaleup, speedup, sizeup and is shown that the proposed algorithm performs better.