摘要
基于方差分析、列联表检验以及兴趣度的定义,提出一种挖掘水文时间序列关联规则优化算法。算法把水文时间序列数据属性分成条件属性和决策属性,通过方差分析和列联表检验在关联规则生成之前剔除的属性和属性值;同时根据新的兴趣度定义,发现"有趣"规则。实验结果证明算法在水文时间序列分析的可行性。
An optimized algorithm for mining association rules in hydrological time series is proposed on the foundation of the analysis of variance (ANOVA), contingency table test and the new definition of interestingness. The data in hydrological time series is divided into condition attribute and decision attribute, then the irrelevant attributes and values of attributes can be eliminated before the generation of the rules using the ANOVA and contingency table test; meanwhile, interesting rules can be generated with the new definition of interestingness. The results confirm the feasibility of the algorithm in the analysis of hydrological time series.
出处
《微电子学与计算机》
CSCD
北大核心
2007年第10期126-129,共4页
Microelectronics & Computer
基金
国家水利部948项目(200517)
关键词
时间序列
关联规则
离散化
统计分析
兴趣度
time series
association rule
discretization
statistic analyst
interestingness