摘要： 在数据挖掘的关联规则挖掘算法中，传统的频繁模式挖掘算法需要用户指定项集的最小支持度。引入Top-k模式挖掘概念的改进算法虽然无需指定最小支持度，但仍需指定阈值k。针对上述问题，对传统挖掘算法进行改进，提出一种新的频繁模式挖掘算法(TNFP- growth)。该算法无需指定最小支持度或阈值，按照支持度降序排列进行模式挖掘，有序地返回频繁模式给用户。实验结果证明，该算法的执行效率更高，具有更强的伸缩性。
关键词:
数据挖掘,
关联规则,
频繁项集,
频繁模式,
Top-k模式
Abstract: In association rule mining algorithm of data mining, conventional frequent pattern mining algorithms need the user to specify minimum support of itemsets. By using the Top-k pattern mining concept, some algorithms are improved to need no minimum support. However, threshold k need be specified. Based on all above, a new algorithm is proposed, which is called TNFP-growth. The algorithm need not specify minimum support or thresholds, and mines patterns with the descending order of their support values and returns frequent patterns to users sequentially. Experimental result proves that it has a high executing efficiency and good scalability.
Key words:
data mining,
association rule,
frequent itemsets,
frequent pattern,
Top-k pattern
中图分类号:
神鹏飞, 王希武, 耿志广, 王创伟, 李国良. 一种无阈值的频繁模式生成算法[J]. 计算机工程, 2011, 37(8): 31-33.
SHEN Feng-Fei, WANG Xi-Wu, GENG Zhi-An, WANG Chuang-Wei, LI Guo-Liang. Threshold Needless Frequent Pattern Generating Algorithm[J]. Computer Engineering, 2011, 37(8): 31-33.