Multivariate Time Series Imputation by Graph Neural Networks and other methods
Note that, given the size of the files, the datasets are not readily available in the folder. See the next section for the downloading instructions.
All the datasets used in the experiment, except CER-E, are open and can be downloaded from this link. The CER-E dataset can be obtained free of charge for research purposes following the instructions at this link. We recommend storing the downloaded datasets in a folder named datasets
inside this directory.
The config
directory stores all the configuration files used to run the experiment. They are divided into folders, according to the model.
The support code, including the models and the datasets readers, are packed in a python library named lib
. Should you have to change the paths to the datasets location, you have to edit the __init__.py
file of the library.
The scripts used for the experiment in the paper are in the scripts
folder.
-
run_baselines.py
is used to compute the metrics for theMEAN
,KNN
,MF
andMICE
imputation methods. An example of usage ispython ./scripts/run_baselines.py --datasets air36 air --imputers mean knn --k 10 --in-sample True --n-runs 5
-
run_imputation.py
is used to compute the metrics for the deep imputation methods. An example of usage ispython ./scripts/run_imputation.py --config config/grin/air36.yaml --in-sample False
-
run_synthetic.py
is used for the experiments on the synthetic datasets. An example of usage ispython ./scripts/run_synthetic.py --config config/grin/synthetic.yaml --static-adj False
We run all the experiments in python 3.8
, see requirements.txt
for the list of pip
dependencies.