Implementation of "Evolutionary Community Detection in Dynamic Social Networks".

If you find this method helpful for your research, please cite this paper.

@INPROCEEDINGS{8852006, 
	author = {Fanzhen Liu and Jia Wu and Chuan Zhou and Jian Yang},
	booktitle = {2019 International Joint Conference on Neural Networks (IJCNN)},
	title = {Evolutionary Community Detection in Dynamic Social Networks},
	year = {2019},
	pages = {1-7},
	doi = {10.1109/IJCNN.2019.8852006}
}

Requirement

• Matlab >= 2013a

Datasets

Datasets used in this paper can be obtain from the original sources.

How to use

Before run the run_ECD.m, please choose a network to load and properly set corresponding parameters in the run_ECD.m. Please see the content in run_ECD.m for more information.

The W_Cube.mat records the adjacent matrices of a dynamic network in several time steps, the GT_Cube.mat or GT_Matrix.mat records the ground truth community structures of a synthetic network in all time steps. For Real-world networks without ground truth community structures, the firststep_DYNMOGA_cell.mat and firststep_DYNMOGA_enron.mat record the community structures detected by the first step of DYNMOGA (Folino and Pizzuti 2014) in all timesteps as the ground truth community structures.

Finally, ECD_Result records the community structures detected by ECD in all time steps; DynMod and DynNmi record the modularity of the detected community structure at each time step and the NMI that measures the similarity between a detected community structure and the ground truth at each time step, respectively.

Disclaimer

If you find any bugs, please report them to me.