利用抽样技术分布式开采可变精度的关联规则

DISTRIBUTED MINING ADJUSTABLE ACCURACY ASSOCIATION RULES USING SAMPLING

关联规则是数据开采的重要研究内容 .利用抽样及元学习技术提出一种快速的分布式开采可变精度的关联规则算法 .为了能获得更准确的结果 ,还给出采用适当缩小最小支持度和扩大全局检测的候选项集等技术的若干改进算法 .最后给出这种方法与类似方法的比较情况 .算法具有效率高和通信量小的特点 ,尤适合于效率比准确性要求更高的场合 .

Association rule mining is an important task of data mining. A fast distributed algorithm for mining adjustable accuracy association rules is presented using sampling and meta learning. In order to acquire more complete results, several variants of the algorithm are also discussed by selecting smaller minimum support and extending global candidate itemsets. The method is compared with similar algorithms. The algorithm is more efficient and has less communicated loads, applicable to those applications where the efficiency could be more important than accuracy results.