search

suanrong / SDNE

This is a implementation of SDNE (Structural Deep Network embedding)
http://www.kdd.org/kdd2016/subtopic/view/structural-deep-network-embedding
322 stars 119 forks source link

Is blogCatalog_embedding.mat training on the optimized parameters? #5

Closed jm-huang closed 6 years ago

jm-huang commented 6 years ago

I run check_multi_label_classification() on blogCatalog_embedding.mat, but i can't get the results shown in paper. Is that i doing something wrong?

My code logic:

  1. load embeddingResult/blogCatalog_embedding.mat using scipy.io
  2. load node labels from file GraphData/blogCatalog3-groups.txt
  3. feed node embedding and labels into function check_multi_label_classification()

the results is about 10 points less than reported in the paper.

suanrong commented 6 years ago

You can retrain the model. The blogCatalog_embedding.mat may be a good result for network reconstruction, not for classification. I am not sure about that.