Galaxy Network Embedding: A Hierarchical Community Structure Preserving Approach

Authors: Lun Du, Zhicong Lu, Yun Wang, Guojie Song, Yiming Wang, Wei Chen

How to use

• We upload a sample dataset Hamilton and you can run our algorithm on this dataset by the following command:

python main.py --conf hamilton_real2

• After running the command above, you can find the node embedding and classification results in ./res.

Introduction of Data Structure

• "edges_hamilton.txt" saves the structure of network.
• "flag_hamilton.txt" saves the labels of nodes that are utilized as ground truth in the node classification task.
• "tree2_hamilton" saves all the edges of the tree, and each pair number is a directed edge including the predecessor node ID in first and the successor node ID in back. And the leaf of the tree represents the original node in the network and their IDs are consistent.