-
The magnetic laplacian provides an alternative to symmetrizing the adjacency matrix for directed graphs. See e.g. "Graph Signal Processing for Directed Graphs based on the Hermitian Laplacian".
-
### Describe the bug
Unsupervised GraphSAGE works by creating context pairs of node IDs using `UnsupervisedSampler` and essentially using them as links for a link prediction task. One can try unsup…
-
Hi,
I'm encountering "[WARNING](pygsp.graphs.graph.check_weights): The main diagonal of the weight matrix is not 0!" when running GraphWave.fit(G)
When passed in to .fit() my graph has a 0 diagona…
-
We have a billion node business graph and want to try RolX for role discovery. How large can RolX scale? We could partition our graph if we need to, but we need to know the limits of RolX.
-
Observing https://github.com/jboynyc/textnets/blob/trunk/docs/tutorial.rst it uses the Degree-Closeness-Betweenness Trifecta for explaining the term and topic significance.
For example, for Network…
-
Hi how do I install the package. I was not able to install it via pip. I tried cloning the repo into a folder but when I import graphwave it is having issues importing.
-
compute_fourier_basis uses self._e, self._U = np.linalg.eigh(self.L.toarray()) this could be replaced with scipy.linalg.eigh(self.L.toarray(), overwrite_a=True) to reuse the memory used by the output …
-
In my recent learning, I'm unsure how to caculate $\lambda_l$. In other words, Do you use test dataset nodes in your calculations of label smoothness? If the situation in the traing set is: The two …
-
Hello, thank you for contributing GraphWave code. I run this code and notice that this is only the process of learning the node representations, and does not evaluate performance. However, I try to do…
-
### Description
Create a blog post for GraphWave (potential promotional material for the 1.0 release) - likely showing a worked example and some background.
### Done Checklist
- [x] KR + LK finalis…