Mining frequent correlated graphs with a new measure
作者：
Highlights：
• A new measure is proposed to capture more interesting inherent correlations in graph databases.
• Downward closure property of the measure achieves faster mining by pruning several candidates.
• Proposed algorithm efficiently mines correlation by building a hierarchical reduced search space.
• Detailed descriptions with real-life examples are given to explain the usefulness of our approach.
• Extensive performance study shows the efficiency, scalability and effectiveness of the algorithm.
摘要
•A new measure is proposed to capture more interesting inherent correlations in graph databases.•Downward closure property of the measure achieves faster mining by pruning several candidates.•Proposed algorithm efficiently mines correlation by building a hierarchical reduced search space.•Detailed descriptions with real-life examples are given to explain the usefulness of our approach.•Extensive performance study shows the efficiency, scalability and effectiveness of the algorithm.
论文关键词：Data mining,Knowledge discovery,Correlated patterns,Graph mining
论文评审过程：Available online 2 September 2013.
论文官网地址：https://doi.org/10.1016/j.eswa.2013.08.082
