Searching of interesting itemsets for negative association rules

Abstract: In this paper, we propose an algorithm of searching for both positive and negative itemsets of interest which should be given at the first stage for positive and negative association rules mining. Traditional association rule mining algorithms extract positive association rules based on frequent itemsets, for which the frequent itemsets, i.e. only positive itemsets of interest are searched. Further, there are useful itemsets among the frequent itemsets pruned from the traditional algorithms to reduce the search space, for mining of negative association rules. Therefore, the traditional algorithms have not come true to find negative itemsets needed in mining of negative association rules. Our new algorithm to search for both positive and negative itemsets of interest prepares preconditions for mining of all positive and negative association rules.