DynaMo: Dynamic Community Detection by Incrementally Maximizing Modularity

A preprint version of our paper: https://arxiv.org/abs/1709.08350.
This is a sample code of our DynaMo paper, including all the necessary implementations for our experiments on the synthetic dynamic networks: 1) synthetic_exp_1.java: using the ground truth communities as the initial community structure; 2) synthetic_exp_2.java: using the result of static algorithm (i.e., Louvain) as the initial community structure.
The necessary implementations for our experiments on the real-world dynamic networks is in real_world_network_exp.java. The real-world dynamic network datasets will be added soon.

We use RDyn to generate the synthetic dynamic networks. [RDyn: graph benchmark handling community dynamics.] (https://github.com/GiulioRossetti/RDyn/blob/master/README.md)
We use # xmeasures to evaluate the performance of the community detection results. [xmeasures - Extrinsic Clustering Measures] (https://github.com/eXascaleInfolab/xmeasures)

Download or clone the whole repository.
Import our code as a Maven project into Eclipse.
Go to the folder ./xmeasures-master and build xmeasures using make release.
If you are using Python 2.7, nothing need to be changed. If you are using Python 3.XX, changing python RDyn-master/rdyn in synthetic_exp_1.java and synthetic_exp_2.java to python3 RDyn-master/rdyn.
Run synthetic_exp_1.java and synthetic_exp_2.java for the synthetic dynamic network experiments.
Run real_world_network_exp.java for the real-world dynamic network experiments. (datasets to be added soon.)

Contact info: Di Zhuang - ‬zhuangdi1990@gmail.com

duanmuxiangyu/dynamo