sam1224 / stcpa Goto Github PK

Hello, I would like to ask some questions about a few files in the newyork data set you provided:
newyork_train_x.npy,
newyork_train_y.npy,
newyork_train_m.npy
are these files represent the original data set? And x represents the original incomplete data, y represents the original complete data, m represents mask matrix?
Or are these files processed after you have fixed the missing rate of 0.5? Is the missing method MCAR(missing completely at random)?
Looking forward to your reply, thank you!

Hi, I take the liberty to disturb you. I see that your article uses ST-SCL（KDD2021） as your baseline. Would you like to share the source code of the reproduction？Thank you！

Hi, how did you pre-train the parameters of your initial GCN model?