How to process the original data into similarity matrix?

HKUST-KnowComp / FMG

KDD17_FMG

138 stars 55 forks source link

How to process the original data into similarity matrix? #4

Closed liangqq6 closed 6 years ago

liangqq6 commented 6 years ago

May I ask how to process the original data into similarity matrix? It seems that the file "mf_features_generator.py" is used to conduct matrix factorization from each similarity matrix corresponding to each meta graph. But how to get the similarity first? That is, how to implement section 2.1 in the paper? Looking forward to your reply, thanks!

hzhaoaf commented 6 years ago

@liangqq6 Sorry for the late reply. I updated the code to compute the meta-graph based similarity matrices. You may check the latest ReadMe file.