Application of Selected Models and Methods of Graph Mining on the Example of the Analysis of Publication Links

The paper deals with specific group of methods of graph and network mining related to the so-called community detection (clustering). Presented methods was discussed and divided into four groups: node-centric, group-centric, network-centric, hierarchical-centric. The characteristics and properties of several methods from each group were discussed. In many cases, attention has been drawn to the computational complexity of the methods presented and ways of improving it. An extensive example of the use of selected methods for the analysis of publication cooperation of employees of a research institute is presented.