Apriori算法低频规则的有效性及实现

Effectiveness and implementation of low frequency rule based on Apriori algorithm

针对经典Apriori算法基于全局、高频两个条件的缺陷,指出事务数据库低频规则的有效性,并通过对C4.5决策树的规则构造,进一步证明事务数据库存在低频规则,在此基础上,给出了一种Apriori低频规则挖掘算法。该算法与经典的Apriori算法兼容,但不是对Apriori算法简单的扩展,而是从理论上打破了Apriori算法基于全局和高频两个条件。最后通过实例用Apriori低频规则挖掘算法和C4.5算法对实例数据库进行挖掘,证明两者的一致性和Apriori低频规则的有效性,同时也证明了Apriori低频规则挖掘算法的有效性。

Firstly,the defects of classical Apriori algorithm based on global view and high frequency were pointed out,and the effectiveness of low frequency rule of transaction database was presented.By constructing the rules of C4.5 decision tree,that the low frequency rule exists in transaction database also was proved.On the foundation of this,a mining algorithm based on low frequency rule of Apriori algorithm was given,which was compatible with classical Apriori algorithm.However,it was not a simple extension of Apriori algorithm,it had broken theoretically Apriori algorithm view based on global view and high frequency.Finally,case database was mined by mining algorithm based on low frequency rule of Apriori and C4.5 algorithms,and the consistency of two methods and the effectiveness of low frequency rule were proved.Moreover,the effectiveness of mining algorithm based on low frequency rule of Apriori algorithm was validated.