Closed Frequent Subgraph Mining by Graph Closure Operation and Its Parallelization

在这项工作中，我们提出了图闭包算子，以及一种利用图模式出现的有效图同构检查方法。给定一个子图，该算子生成一个包含该子图的闭子图，其中闭子图是与给定子图频率相同的子图中的最大子图。然后，我们开发了一种高效的方法来枚举图数据库中所有可能的闭频繁子图，并将其内部进程并行化以提高其可扩展性。

In this work, we proposed the graph closure operator, as well as an efficient method for graph isomorphism checking that takes advantage of occurrences of a graph pattern. Given a subgraph, the operator generates a closed subgraph including it, where the closed subgraph is the maximum subgraph among those which have the same frequency as the given one. Then, we developed an efficient method for enumerating all possible closed frequent subgraphs in a graph database by means of them, and parallelized its inner processes to improve its scalability.