The essence of algorithms based on the graph theory is to represent target objects in graph form. Graph vertices correspond to objects, and the edge weights are equal to the distance between vertices. The advantages of graph clustering algorithms are their excellent visibility, relative ease of implementation, and their ability to make various improvements based on geometrical considerations. The main algorithms based on graph theory are the algorithm for selecting connected components, the algorithm for constructing the minimum spanning tree, and the multilayer clustering algorithm.
The algorithm for selecting connected components is based on the R input parameter, and the algorithm removes all edges ...
