Our algorithms are implemented in Python 3.10.12 and all experiments are executed on a server with an Intel (R) Xeon (R) E5-2680 [email protected] CPU and 256GB RAM running Ubuntu 18.04.
Ps: Our code can also run on a desktop with Apple M1 and 8GB RAM running macOS Monterey 12.3 (and Inter(R) Core(TM) [email protected] and 16 GB RAM running Windows 10). But we recommend you run it on a server because the server has enough memory to handle large datasets and run faster.
We focus on identifying the communities from a temporal network, in which each temporal edge is associated with a timestamp. In particular, temporal edges are stored in the raw data where each line is one temporal edge.
| from_id | \t | to_id | \t | timestamps |
|---|
Due to the space limit, we only upload some small datasets. But, you can download all original datasets used in our paper from the following table or the preprocessed datasets from
https://www.dropbox.com/scl/fo/90casjr51m85wr5l5duhm/h?rlkey=zvgyxhhxxu4qvq6c4iiqvxp5p&dl=0.
If you have any questions, please contact [email protected]
You may use git to clone the repository from GitHub and run it manually like this
git clone https://github.com/longlonglin/QTCS.git
cd QTCS
python qtcs.py data/Facebook
The running results are as follows
data/Facebook is loading...
loading_graph_time(s)4.507202625274658
number of nodes: 45813
number of static edges: 183412.0
number of temporal edges: 585743.0
number of timestamp: 1473
self.tmax:552
compute_ttp_time(s)113.53539848327637
seed25115
time_tppr(s)2.2104275226593018
egr_time(s)3.710035562515259
time_expanding(s)0.10440444946289062
time_reducing(s)0.013033390045166016
seed8401
time_tppr(s)83.49688935279846
egr_time(s)85.08008170127869
time_expanding(s)39.289947748184204
time_reducing(s)0.5261859893798828
seed16973
time_tppr(s)2.932955265045166
egr_time(s)5.026332855224609
time_expanding(s)1.946237325668335
time_reducing(s)0.24632477760314941
seed38625
time_tppr(s)20.524596691131592
egr_time(s)22.073683977127075
time_expanding(s)2.5056912899017334
time_reducing(s)0.0404210090637207
seed29551
time_tppr(s)1.3863840103149414
egr_time(s)2.8778045177459717
time_expanding(s)0.025407791137695312
time_reducing(s)0.0007688999176025391
Our model has only one parameter, alpha, which ranges from 0 to 1, and its default value is 0.2. If you want to change alpha, you can modify it in line 553 of qtcs.py.