search

benedekrozemberczki / GraphWaveletNeuralNetwork

A PyTorch implementation of "Graph Wavelet Neural Network" (ICLR 2019)
GNU General Public License v3.0
576 stars 116 forks source link

what's the meanning of the "feature matrix"? #7

Closed CindyTing closed 3 years ago

CindyTing commented 4 years ago

Hello author, sorry about a stupid question. But the Cora dataset has Cora.cites corresponding your cora_edges.csv, and Cora.content's paper index and paper category for your cora_target.csv, so I don't understand the meanning of your cora_features.json . In the beginning, I just think it's an adjacency matrix of all nodes(paper index), however, the content are inconsistent. Such as ,in cora_edges.csv it's as the picture as follw: image and in cora_features.json it's : image So I am confused , and hope for your answer. Thank you very much.

CindyTing commented 4 years ago

I understand, and I am so stupid, that's the index of bag of words. I am so sorry, thanks anyway.

tanjia123456 commented 4 years ago

hello, could you plaese explain cora_feature.json for me? I use this in my own datasets, so i'm not understant the meanning of this format.

CindyTing commented 4 years ago

@tanjia123456 There is a dictionary composed of the word of all documents, each line of cora_feature.json represents the index of the word (in this document) in the dictionary.

tanjia123456 commented 4 years ago

@CindyTing Can I understand it as the feature index corresponding to each node?

CindyTing commented 4 years ago

@tanjia123456 yeah, I think so.

tanjia123456 commented 4 years ago

ok, thank you very much. I will try to make data in this format

benedekrozemberczki commented 3 years ago

Yes it is a sparse matrix in row column index format.