Skip to content

premolab/GraphEmbeddings

Repository files navigation

GraphEmbeddings

This repository contains realization of DDoS (aka Histogram loss) algorithm for generating graph embeddings.

Structure

All code is stored in folder final_src

Realization of several embedding algorithm including our algorithm can be found in folder transformers.

Folder io_utils contains code responsible for reading and writing graphs and embeddings.

Folder transformation contains generic code to generate an embedding with any available algorithm.

Other folders represent sets of experiments for comparing algorithms: link_prediction, classification and clusterization.

How to run

To make this code work you need to replace path to necessary dataset in file final_src/settings.py with your local path.

About

Python implementation of the DDoS (“Histogram loss”) graph node embedding algorithm, with experiment pipelines for link prediction, node classification, and clustering. Paper: https://arxiv.org/abs/1810.03032

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages