Collective Causal Discovery Algorithm for Multivariate Time Series

面向多元时间序列的群体因果关系发现算法,同时也是一种基于因果结构的多元时间序列聚类算法

代码实现基于 pytorch 1.71

General setting: number of variables is 5; Erdos-Renyi model with parameter 0.3; lag is 1.

Run a demo of Our model :

python main.py

and the baseline methods :

python baseline/baseline_main.py

Reproduce synthetic experiments:

Our model :

nohup bash run.sh &

the results will be stored in 'output' folder and to summarize the results with all settings :

python result_combine.py

the results under different settings will be stored in 'result' folder;

And the baseline methods :

nohup bash baseline/baseline_run.sh &

the results will be stored in 'baseline/baseline_output' folder, and to summarize the results with all settings :

cd baseline
python baseline_result_combine.py

the results under different settings will be stored in 'baseline_result' folder.