Example of config file for visuaization after inductive prediction

[clairelaudy]

Hi,
Could it be possible to have an example of the config file needed to visualize the predictions and exlapnations after using the eval_and_predict_inductive.py script ?
The one provided for the non inductive (config/mockdata_viz.yaml) example doesn't seem to apply.

It generates a ValueError:

File "/opt/gensoft/exe/BioPathNet/0.1/libexec/visualize.py", line 124, in
train_set, valid_set, test_set = _dataset.split()
ValueError: not enough values to unpack (expected 3, got 2)

[clairelaudy]

The predictions.csv file also contains negative values as probabilities for some triples, I don't understand how this should be interpreted.

[emyyue]

Check out this file:

BioPathNet/config/mock/mockdata_inductive_pred.yaml

Line 1 in c6f7afc

output_dir: experiments/

Regarding negative probabilities:

• Common KG metrics are ranking based. So that the neg scores do not matter. If you work with probablities, then take the sigmoid. The naming of the column is unfortunate.

[emyyue]

Commit b051a90 renames column from 'probability' to 'prediction_score'