search

THUDM / GRAND

Source code and dataset of the NeurIPS 2020 paper "Graph Random Neural Network for Semi-Supervised Learning on Graphs"
MIT License
206 stars 40 forks source link

Normalization on X #5

Open syxd opened 3 years ago

syxd commented 3 years ago

Why apply row normalization on the raw node feature before training? This procedure is not mentioned in the paper. And I find if we remove this normalization procedure for data preparation, the performance will changes a lot.

wzfhaha commented 3 years ago

This normalization strategy is used in almost all GNN models for this task. Here we keep the same configuration as the previous works to make sure the comparisons to be fair.