Frequent Subgraph Mining in Graph Databases Based on MapReduce

Abstract

In recent years, graph mining has become a popular research direction in the area of data mining. Frequent subgraph mining is an important technology of graph mining that can be used in many fields such as chemical informatics, bioinformatics, and social sciences. The increasing size of graph database is challenging traditional methods of subgraph mining. In this paper, we propose a new approach based on MapReduce to mine frequent subgraph patterns from the vertex-classified graph databases in large sizes. There are two rounds operation to MapReduce. The first round is to mine the locally frequent subgraphs in each node and then we collect the results for all nodes and filter some redundant graphs to obtain a set of frequent subgraphs candidate in global view. The second round is to calculate the global frequency for each graph using the set of candidate generated by the first round. Some topical frequent subgraphs are filtered according to special requirement. The experimental results show that this approach reduces the execution time when dealing with large graph databases.