【摘要】 以往算法的研究主要围绕着减少候选项目集进而减少事务数据库的扫描次数的角度,先求出候选项集,再计算候选项集的支持度求得频繁项集。本文改变过去求频繁项集的角度,从新的角度来看频繁项目集的定义,同时结合人工免疫的特点,设计一个基于人工免疫的新频繁项集挖掘算法。本文详细介绍了算法设计等。新算法的复杂度与支持度,数据库总容量有关。验证实验的结果与其他算法相比较证明了该算法的可行性、有效性和完备性。更多还原

【Abstract】 Tradition algorithm attempted to improve the mining efficiency reducing the number of database passes to control the I/O cost, which at first derives candidate itemsets from tuples in database, and count support of candidate itemsets to get frequent itemsets. Changing the former way of discovering frequent itemsets, we can understand the defi- nition of frequent itemsets from another standpoint. Making use of atifical immune strategy, this paper present a new algorithm. In the paper, the algorithm is introduced in detail. The cost of new algorithm is related to the support and the total number of database. Finally, simulation shows the correctness and validity compared with other algorithm.更多还原