基于质数理论的最大频繁项集挖掘研究

Research on Mining Maximum Frequent Itemsets Based on Prime Number Theory

利用质数的特性,采用质数积代替事务将事务,数据库转换成质数积数据集,采用数据集二维数组保存质数积之间的整除关系和最大公约数信息。PNMax算法利用数据集二维数组可以快速挖掘出最大频繁项集,并且数据集二维数组在挖掘过程中将持续减少所占空间。最后通过实验验证了算法的可行性和优越性。

Using the characteristics of prime number, the transaction database is transformed into a data set of prime product by using the product of prime number instead of transaction, and the two-dimensional array of data set is used to save the integer division relationship and the greatest common divisor information between the products of prime number. PNMax algorithm can quickly mine the maximum frequent itemsets by using this array, and this array will continue to reduce the space occupied in the mining process. Finally, the feasibility and superiority of the algorithm are verified by experiments.