摘要
关联规则挖掘会产生大量的项集和规则,其中只有少部分是用户感兴趣和有价值的,其他大部分是冗余的或已知的。在已有的空间关联规则挖掘研究中,用户对数据库中存在的强制约束缺少考虑。本文分析了空间数据库中的已知空间依赖,发现已有的Apriori算法和闭频繁项集挖掘难以消除该空间依赖,为此提出了基于概念格方法的已知空间依赖剔除策略,包括概念格中每个闭频繁节点的产生子获得方法和利用概念格产生子实现最优频繁地理模式挖掘的方法,最后通过实验验证了概念格产生子方法的有效性和优越性。
Association rule mining could bring a lot of itemsets and rules,only parts of them would attract user＇s interesting and valuable.In the existing studies of spatial association rule mining,researchers do not take the compellent restriction concealed in database into consideration.After introducing the existing compellent restriction,That Apriori and Colse Set mining algorithms can＇t guarantee to eliminate known restriction was pointed out in this paper.Then the relevant concepts of Concept Lattice as well as its generator were analyzed.And beyond the analysis,a way to eliminate known spatial restriction with Concept Lattice were proposed,including the method of generated factors in closed frequent itemsets and the most optimized frequent mining methods by using generator of concept lattice.Finally,an experimentation was taken to validate the validity of the proposed way.
出处
《测绘科学技术学报》
CSCD
北大核心
2013年第3期304-307,共4页
Journal of Geomatics Science and Technology
基金
国家自然科学基金项目(40871183
41140012
41271392)
关键词
空间依赖
概念格
产生子
关联规则挖掘
闭频繁项集
spatial restriction
Concept Lattice
generated factors
association rule mining
closed frequent itemset
