Node Attribute Prediction

Replication code for ICWSM'21 Paper ``What is the Node Attribute Prediction Task? Within-Network, Across-Network, versus Across-Layer Tasks''.

Documentation

This repository contains all the corresponding code to recreate tables and figures.

Directions

1. Save data in corresponding data folders.

2. How to run code where folder organization is briefly described below:

• ./data/: Placeholder folder for storing raw datasets. We use the Facebook100 (FB100) dataset (Traud et al. 2011; Traud, Mucha, and Porter 2012), the Copenhagen Networks Study (Stopczynski et al. 2014) and the Reality Mining dataset (Eagle and Pentland 2006).
• ./code/: Code for recreating tables and figures in main paper. Note that relative file paths will have to be updated in the code as needed.
• ./figs/: Folder for all figures.

3. For all the ReFeX code, we adapt from https://github.com/LLNL/refex-rolx.

Authors

• Kristen M. Altenburger, kristenaltenburger@gmail.com
• Johan Ugander, jugander@stanford.edu