A distributed weakly connected component solving algorithm based on dyeing

A distributed weakly connected component solving algorithm based on dyeing

Shaoyou Tan Juxian Li Pu

The data of link relation such as social network and bioinformatics Network is exploding, the demand of large graph processing is more and more urgent, the traditional graph algorithm dependence on single point of memory can't meet the demand of rapid growth of data scale. In this paper, a fast and extensible iterative dyeing algorithm CR is proposed for solving the problem of weak connected components of graphs, and the MapReduce model of the algorithm is established. Finally, on the Hadoop platform, the four sets of communications data provided by the Stanford University social Networking Analysis laboratory were tested and compared with the algorithms provided by the open source Data Mining toolbox Xrime.

A distributed weakly connected component solving algorithm based on dyeing