Extracting Schemas from Large Graphs with Utility Function and Parallelization
Extracting Schemas from Large Graphs with Utility Function and Parallelization
复制标题
使用效用函数和并行化从大图中提取模式
DOI：
10.1007/978-3-319-91455-8_13
复制
发表时间：
2018
期刊：
影响因子：
--
通讯作者：
Y. Sekine and N. Suzuki
中科院分区：
文献类型：
--
作者：
赤澤豪樹，鈴木伸崇;Y. Tsuboi and N. Suzuki;Y. Sekine and N. Suzuki
与关系数据库和XML文档不同，大多数图都没有给出自己的模式。如果我们能有效地从图中抽取模式，我们就可以利用抽取的模式进行查询优化、结构浏览等。虽然效用函数可以提取合理的模式，但效用函数的主要问题是它的计算成本。在本文中，我们提出了一个模式提取算法的基础上（a）一个新的效用函数称为本地效用函数和（B）并行化。实验结果表明，该算法可以在不损失模式质量的前提下，更有效地从图中提取模式。
Unlike relational databases and XML documents, most of graphs are not given their own schemas. If we can extract a schema from a graph efficiently, we can take advantage of the extracted schema for query optimization, structure browsing, and so on. In this paper, we consider extracting schemas from large graphs by usingutility function. Although reasonable schemas can be extracted by the utility function, the major problem of the utility function is its computation cost. In this paper, we propose a schema extraction algorithm based on (a) a novel utility function called local utility function and (b) parallelization. Experimental results show that our algorithm can extract schemas from graphs more efficiently without losing quality of schemas.