search

Aaditya-Singh / E2E-ECPE

Code for the WASSA 2021 paper (oral) "An End-to-End Network for Emotion-Cause Pair Extraction"
23 stars 7 forks source link

dataset #1

Closed wjczf123 closed 3 years ago

wjczf123 commented 3 years ago

I have a qustion about dataset. I found that the corpus consists of 2843 documents. But Gao et al. writed that this corpus has 2403 documents (Table 1 in their paper). And Li et al. and Fan et al. writed that this corpus has 2156 documents (see Table 3 and Table 1 in their paper). What caused this difference?

shreeshail-23 commented 3 years ago

We could not find the dataset publicly since the web-links we found in previous publications were expired. We obtained the dataset from Prof. Lin Gui, University of Warwick, and it contained 2843 documents.

wjczf123 commented 3 years ago

OK. Thank you a lot.