International Journal of Computer Science Issues

More than a traditional journal...

Most association rule mining approaches rely on predetermined support and confidence values to find certain relationships among database itemsets. However, specifying minimum support and confidence values of the mined rules in advance often leads to either too many or too few rules, which negatively affects the performance of the overall system. To address this issue, this paper presents a new algorithm based on null hypothesis to find non-coincidental relations among different itemsets in large databases without a prior defined threshold values. Intensive simulated experiments have been performed on different databases to confirm the validity of the suggested algorithm. The obtained results show that there is a significant improvement in the system performance in terms of the number of frequent items used, the number of generated rules, and the run time.