Not sure the right terminology, but I am working on a node classification problem, and I am only able to obtain a training accuracy of 50% for my GNN (test accuracy is 40%).

It seems I should be able to overfit the model and achieve nearly 100% with the training data, but I assume the feature vectors are not adequate for what I am trying to learn with the model.

How could I verify this is the case? Is there an approach for determining if the feature space is adequate for the problem?



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.