Mining multiple level association rules under weighted concise support framework

Association rules tell us interesting relationships between different items in transaction database. Traditional association rule has two disadvantages. Firstly, it assumes every two items have same significance in database, which is unreasonable in many real applications and usually leads to incorrect results. Secondly, traditional association rule representation contains too much redundancy which makes it difficult to be mined and used. This paper addresses the problem of mining weighted concise association rules based on closed itemsets under weighted support-significant framework, in which each item with different significance is assigned different weight. Through exploiting specific technique, the proposed algorithm can mine all weighted concise association rules while duplicate weighted itemset search space is pruned. As illustrated in experiments, the proposed method leads to better results and achieves better performance.