计算机科学
图形数据库
距离遗传图
子图同构问题
图形
因子临界图
图因式分解
搜索引擎索引
数据挖掘
聚类分析
理论计算机科学
折线图
图形功率
情报检索
人工智能
作者
Fathimabi Shaik,Ebenezer Jangam,R. B. V. Subramanyam,A. Srisaila
标识
DOI:10.23919/indiacom54597.2022.9763250
摘要
In a distributed environment, as the graph databases emerge from multiple sources, the graph database's size increases rapidly, resulting in a massive database. Mining frequent subgraphs is a fundamental problem for graph clustering, classification, and subgraph association rule mining. Nevertheless, extracting frequent subgraphs from an extensive graph database is complex work. Existing works for frequent mining subgraphs assume that the large graph database is stored on a single system. This assumption is not correct in all cases. This paper designs a novel algorithm in which frequent subgraph mining takes advantage of partitioning and graph indexing. We follow three necessary steps: graph partitioning, graph indexing, and frequent subgraph mining. We used structure-based graph partitioning for frequent subgraph mining. We use Integrated Graph Index as a vertical format of graphs in the case of frequent itemset mining. Integrated Graph Index has low maintenance costs, and it provides an easy way to add graphs in the massive graph database. We complete frequent mining subgraphs in a single Map-Reduce round. This approach is suitable for incremental mining of frequent subgraphs from the large graph database. We proved that this technique reduces running time to find frequent subgraphs over large graph databases.
科研通智能强力驱动
Strongly Powered by AbleSci AI