Mastering Machine Learning with Spark 2.x

Multigraphs

Generally, a graph without loops or multiple edges is called simple. Most graphs we will encounter in the applications of this chapter do not share this property. Very often, graphs constructed from real-world data will have multiple edges between vertices. In literature, graphs with multiple edges are referred to as multi-graphs or pseudo graphs. Throughout the chapter, we will stick with the multigraph notion and will follow the convention that such a multigraph can include loops as well. Since Spark supports multigraphs (including loops), this notion will be very useful in the applications. In the following diagram, we see a complex multigraph with multiple connected components:

Figure 8: A slightly more involved social ...