About statistics of News14

[kimwongyuda]

Thank you for your nice work.

I compute # of documents and # of stories by using dataset from https://github.com/Priberam/news-clustering.

The number of documents is same with that in paper, 16136.
However, the number of stories is different with that in paper, 733 vs 788.
I counted unique values of "cluster" keys from the dataset.

Also, since dev set and test set in the dataset are time-independent, if there are same cluster values between dev set and testset, i regard them as distinct values.

But there are just 7 overlapping cluster values.

Could you tell me how you counted the number of stories?