水平分布数据库的正负关联规则挖掘

Positive and Negative Association Rules Mining on Horizontally Partitioned Database

目前,正负关联规则的挖掘受到越来越多的关注,在现实运用中也越来越广泛。随着信息技术和经济全球化的发展,许多企业都在全国甚至全世界范围内拥有自己的数据中心。企业通过分析这些数据的关联来为他们的战略抉择和政策制定服务。而直接在这些庞大的数据中寻找数据之间的关联不是一件容易的事。而且,在这些大量的数据中,不是所有的数据都是分析中所需要的。文中通过在文献[4]中所提的方法中引入对不同数据库赋予不同权重值的方式,使得在分布式数据库中挖掘正负关联规则更加高效。经过测试,这一改进是有效的。

Recently,positive and negative association rules mining has received some attention and been proved to be useful in real world.As the developing of information technology and economy＇s globalization,many enterprises and companies have established their data centers around the country,even around the world.They usually mine useful information in their distributed databases to help their decisions making and policies establishing.But,there are thousands of items in each database and it will be expensive if directly mining association rules in such a large database.What＇s more,most of the items are not interesting.Extend the algorithm mentioned in reference[4] by place different weight on different database to make it more efficient to mine both positive and negative association on horizontally partitioned data.Through testing,this optimization is effective.