A MINING ALGORITHM FOR ASSOCIATION RULES BASED ON INTEREST MEASURE

Association rules is one of the main research fields in data mining, and to gain the more valuable rules is a focus in this field. But most of the most current algorithms can only find out the positive-item rules and the interest measure can only be used as a filter. To solve this problem, an idea to find out the negative-item-contained rules is forword put through modification of the related definition and through the introduction of the negative itemset, based on the previous work. The algorithm based on this idea is also given. Finally, the algorithm is evaluated and analyzed through experiments and practices.