The total num of the paper for recom is about 36k and is too big for me. I just need about 10% ~ 20%. Choosing data randomly seems not good.
How can i divide the data and keep citations wherever possible? Do you have any suggestions?
Sorry to say that we've never considered this use case. SciDocs is intended to be used as is, or otherwise it is not comparable to other published results.
The total num of the paper for recom is about 36k and is too big for me. I just need about 10% ~ 20%. Choosing data randomly seems not good. How can i divide the data and keep citations wherever possible? Do you have any suggestions?
Looking forward to your reply and thanks a lot.