create graph data for your test in GNN
The idea came about while experimenting with graph convolutional networks. Graph convolution has received extensive attention in recent years, But in practice, there are always a few graph data for use. And in some specific scenarios, such as: project demonstration, it is often necessary to perform function demonstration for user application scenarios.In addition to the leading companies having their own data sources, most of them may still need to find industry data for experiments.But field-specific data is hard to find, and may not meet the needs of the experiment.In this case, modificate the data provided by the existing framework of GNN, sunch as dgl,may be the only choice. Since it's faked, why don't create your dreaming data for a better Experimental result.
The original intention of this project was to provide data for easier research on GNN. If there are any adverse effects, it has nothing to do with me. I am afraid that the technology is too low-level and will be laughed at, and I want to privatize it -_-.