Users can control the characteristics of generated attributed graphs by acMark.
acMark scales linearly to the number of edges and the number of attributes so massive attributed graphs can be generated.
- numpy == 1.20.2
- scipy == 1.6.2
- powerlaw == 1.4.6
(for demo)
- matplotlab == 3.4.1
- seaborn == 0.11.1
We show the demo in the jupyter notebook
n : number of nodes
m : number of edges
d : number of attributes
k : number of classes
more details in our paper