Mining Weighted Negative Association Rules from Infrequent Itemsets Based on Multiple Supports

The existing researches for association rule mining most use the single minimum support. However, in practical applications, the occurrence frequency of each itemset is different. We should set different minimum support for itemsets. In association rule mining, if the given minimum support is too high, then the items with low frequency of appearance couldn´t be mined. Otherwise, if the given minimum support is too low, then combination explosion may occur. We support the technique that allows the users to specify multiple minimum supports to reflect the natures of the itemsets and their varied frequencies in the database. It is very effective for large databases to use algorithm of association rules based on multiple supports. The existing algorithms are mostly mining positive and negative association rules from frequent itemsets. But the negative association rules from infrequent itemsets are ignored. Furthermore, We set different weighted values for items according to the importance of each item. Based on the above three factors, an algorithm for mining weighted negative association rules from infrequent itemsets based on multiple supports(WNAIIMS) is proposed in this paper.