Distributed implementation of the graphs clusterization goodness metrics calculation using mapreduce and vertex-oriented graph processing model

Abstract

Goal. Research of the implementation features of the graph clusterization goodness metrics using the vertexoriented graph computing model and MapReduce. Materials and methods. The basic concepts of graph theory were used to define the goodness metrics, and the MapReduce with a vertex-oriented graph-computational approach were used to develop the goodness metrics calculation algorithms. Results. Distributed algorithms for calculating graph clustering goodness metrics are proposed and tested. Conclusion. The results can be used to analyze the quality of the partitioning of large graphs that obtained using an arbitrary distributed clustering algorithm.