CAJ | 학술논문

提出了一种基于Rough有限概念格的规则发现方法(RRLIRD)，以揭示数据集中的蕴涵规则.首先引入有限概念格的简化表示形式，由用户选择数据集中感兴趣属性集创建概念格结构，提高用户的交互性和挖掘的效率；然后运用有限概念格与Rough集理论相结合形成Rough有限概念格，蕴涵规则则由其特有的上、下近似运算得到，不需计算繁琐的频繁项目集.算法运用大型超市的交易流水数据进行仿真实验.结果表明，执行时间比经典的Apriori算法大大降低.该算法也适用于证券行情分析和农业数据库中的病虫害分析等.
제출료일충기우Rough유한개념격적규칙발현방법(RRLIRD)，이게시수거집중적온함규칙.수선인입유한개념격적간화표시형식，유용호선택수거집중감흥취속성집창건개념격결구，제고용호적교호성화알굴적효솔；연후운용유한개념격여Rough집이론상결합형성Rough유한개념격，온함규칙칙유기특유적상、하근사운산득도，불수계산번쇄적빈번항목집.산법운용대형초시적교역류수수거진행방진실험.결과표명，집행시간비경전적Apriori산법대대강저.해산법야괄용우증권행정분석화농업수거고중적병충해분석등.
An efficient algorithm was found to discover the implication rules in a data set. As an important data mining technique, the implication rules can help to explore the dependencies among values of attributes of a database. The algorithm first extends the concept lattice theory by building the simplified lattice structure according to the data set with the resticted attributes to improve human interaction and mining efficiency. The constrained concept lattice, together with the rough set theory, is then incorporated into the method to implement a new restricted rough lattice-based implication rules discovery (RRLIRD) approach to interactively acquire the rules with the specific rough upper and lower approximation. The algorithm is different from the classical rule extraction methods without computing the frequent itemsets. For the application to the transaction data set of large-scale supermarkets, a simulation was implemented to demonstrate that the approach can reduce the computational time greatly comparing with that of the Apriori algorithm. The algorithm can also be extended to other areas such as stock analysis and agricultural application.