Hi,
I am comparing OGB and MoleculeNet. On tox21, you reported roc-auc of 77.5% in your paper while roc-auc was 85% with MoleculeNet:
http://moleculenet.ai/latest-results
Even with logistic regression, MoleculeNet reported 78%. Do you know why the results are so different?
The main differences, in my observation, are:
I modified your example (https://github.com/snap-stanford/ogb/blob/master/examples/graphproppred/mol/main_pyg.py), and used random split. Still the ROC-AUC is around 75%. I am now trying to embed node feature as MoleculeNet did. Where can I find the script that prepares the tox21 dataset in OGB? Thanks a lot for your help.
Taro Kiritani
--
You received this message because you are subscribed to a topic in the Google Groups "Open Graph Benchmark" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/open-graph-benchmark/0KML08gNVcM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to open-graph-bench...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-graph-benchmark/4f21f9d8-2f75-429f-bf0a-ca1c30c77630o%40googlegroups.com.
To unsubscribe from this group and all its topics, send an email to open-graph-benchmark+unsub...@googlegroups.com.
To unsubscribe from this group and all its topics, send an email to open-graph-bench...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-graph-benchmark/4f21f9d8-2f75-429f-bf0a-ca1c30c77630o%40googlegroups.com.
--
You received this message because you are subscribed to a topic in the Google Groups "Open Graph Benchmark" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/open-graph-benchmark/0KML08gNVcM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to open-graph-bench...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-graph-benchmark/0fd07ee1-d97e-4de2-bbc8-84716ddc64f4o%40googlegroups.com.