vanzytay / absa_devsplits Goto Github PK

I have question about the polarity label in SE1415, the polarities are numbered by 1, 2, 3, 4.
1 is neg, 2 is pos, 3 is neu
but what is 4 represents? And there are lots of samples that are labeled 4.

Hi, I cannot find the training, dev, test dataset in /data directory. Could you plz upload the split dataset again?

f = open('dataset/Laptop_term.pkl')
info = pickle.load(f)

the error is :EOFError.

But ,other three files are no probles.

As shown in your paper, there should be 499 examples in Dev set of LAPTOP terms dataset, however, I found there are only 328 sentences actually, any ideas? thanks a lot.

Hi，when I look into dataset SE1415, I was confused about the aspect staff. In SE1415, the aspects are indexed into numbers from 1 to 8. But when I try to transfer these index into words by "wordindex" given in the dataset, it seems that the transferes aspects doesn't make sense. So can you tell me which aspect does the aspect index 1-8 stand for, respectively?