search

xdweixia / SGCMC

The Tensorflow Implementation of paper Self-supervised Graph Convolutional Network For Multi-view Clustering
34 stars 3 forks source link

How to use for graph data with one attribute but multiple topologies? #1

Open JLUVicent opened 2 years ago

JLUVicent commented 2 years ago

Hello, I am glad to read your article. I found that the module in your paper is used for graphs with multiple attributes but only one topology. Can your model be applied to graph data with one attribute but multiple topologies? if it can,how should I modify the model? I would appreciate it if you could reply.

xdweixia commented 2 years ago

Thanks for your attention. The released model can also be applied to deal with data with one attribute but multiple topologies. If you have V topologies A^{(1)}, A^{(2)}, ..., A^{(V)}, and one attribute X, then you need to map {X, A^{(1)}}, {X, A^{(2)}}, ..., {X, A^{(v)}} to obtain graph embeddings Z^{(1)}, Z^{(2)}, ..., Z^{(V)} via shared GCN. You can add my we-chat (ID: 15229366656) if you need.