benedekrozemberczki/GraphWaveletNeuralNetwork

Laplacian matrix

Serendipity-LC opened this issue · 4 comments

Hi! Why do we need to use all the node data (2707 nodes) when constructing the Laplacian matrix instead of just the training set data?

Hi! Why do we need to use all the node data (2707 nodes) when constructing the Laplacian matrix instead of just the training set data?

If we do this, do we use the node features of the validation set and test set during the training process?

This is semi supervised.