Year of Graduation
Metaheuristic for Community Detection in Graph
Applied Mathematics and Information Science
In the framework of this paper, an attempt was made to implement a method whose idea is to analyze a list of parameters embedded in the graph structure (for example, the number of vertices and edges, density, etc.) and to determine the most suitable clustering algorithm for the given graph, using one of the classification algorithms. Graphs, for which the true structure of the communities was known, were used for training the classification algorithm, therefore, using certain clustering performance evaluation metrics, it was possible to determine which of the applied algorithms of finding communities shows the most accurate results. The paper dealt only with non-oriented unweighted graphs without loops and multiple edges.