摘要
在进行并行关联规则挖掘时,数据偏斜和工作量平衡这两个数据分布特征影响着剪枝的有效性.本文提出了用定量的方式对数据偏斜和工作量平衡进行度量,并对不同值的组合进行了分析,以便在以后研究算法时可以有效地调整这两个特征值以提高剪枝的性能.
When excavating with parallel association rules, the two data distribution characters, data skewness and workload balance,will affect the validity of pruning. So we bring forward the method of measuring the two char-acters of data skewness and workload balance with the model of fix quantification, and analyze the compages of dif-ferent values, so that we can adjust the two characters efficiently to improve the validity of pruning in the future study of the arithmetic afterwards.
出处
《空军雷达学院学报》
2004年第1期47-49,共3页
Journal of Air Force Radar Academy
关键词
数据偏斜
工作量平衡
分布剪枝
全局剪枝
data skewness
workload balance
distribution pruning
global pruning