模式级链接关联数据集上的关联规则挖掘研究

针对关联数据集合呈现出的大数据特性和蕴含的语义信息,提出了首先建立关联数据集的模式级链接,再进行关联规则挖掘的方法。在同领域RDF数据集上定义RDF数据项模式并提出数据项模式的产生规则;利用RDF数据查询技术从数据项模式获得RDF数据项集合,进而再推导出特定领域内的关联规则。提出的基于关联数据RDF数据项模式的关联规则挖掘方法将关联规则挖掘扩展到同一领域内的数据集合而不再局限于单一数据集,同时给出了基于Hadoop的大规模RDF数据集上的关联规则挖掘的实现方案。实验结果验证了模式级链接对于关联规则挖掘的价值和所提方法的有效性。

Association Rules Mining on Schema-level Interconnected Associated Data

A schema-level interconnected association rules mining method for large scale associated data was proposed based on the semantic information implied in the associated data set.Instead of mining association rules from separated RDF data sets directly, firstly,we established schema-level linkage between different data sets.The RDF data item pattern generation rules are defined based on the schema-level linked datasets and then the RDF data query techniques are exploited for constructing RDF data items sets.The proposed data item patterns generation rules can extend the data mining objects from a single data set to multi-datasets in the same domain.A Hadoop based implementation plan of association rules mining was designed.The experiment results prove the value of establishing schema-level linkage on linked data and the effectiveness of the proposed method.