search

yangheng95 / LCF-ATEPC

codes for paper A Multi-task Learning Model for Chinese-oriented Aspect Polarity Classification and Aspect Term Extraction
MIT License
186 stars 45 forks source link

dataset #25

Closed yassmine-lam closed 3 years ago

yassmine-lam commented 3 years ago

Hi,

could u give us a description about the dataset format?

thank u

yangheng95 commented 3 years ago

For the datasets, each row contains a token or word, an ATE tag and a polarity tag. If a token or word does not belong to an aspect, the polarity is -1, else: 0 for negative， 1 for neutral， 2 for positive.

However, each sample only label polarity for one aspect, other's polarity is -1 the same as non aspect words.

e.g.

I O -1 charge O -1 it O -1 at O -1 night O -1 and O -1 skip O -1 taking O -1 the O -1 cord B-ASP 1 with O -1 me O -1 because O -1 of O -1 the O -1 good O -1 battery B-ASP -1 life I-ASP -1 . O -1

yassmine-lam commented 3 years ago

ok thank u

parasd086 commented 2 years ago

What's the difference between I-ASP and B-ASP?

seominseok48349278 commented 2 years ago

What's the difference between I-ASP and B-ASP?

https://github.com/yangheng95/ABSADatasets/issues/24

liufenghupan commented 6 months ago

您好，我正在构造一个自己的中文数据集，但是看了您的训练数据有一些问题想要请教，我看到您的每一个评论对象当中只标注一个属性，其他属性情感极性标位-1，但是您的所有的评论对象当中都只有一个B-ASP，如果一个评论对象当中有很多属性，每一个属性的标记不是都应该用B-ASP开始吗？还有就是我发现您的中文数据集中的极性只有0和2，为什么中文数据集比英文数据集少了“中性”的情感分类呢？感谢您的工作，希望能得到您的回复，谢谢。

yangheng95 commented 6 months ago

https://github.com/yangheng95/LCF-ATEPC/issues/30#issuecomment-839394501

因为原始中文数据集只有两个类别。

liufenghupan commented 5 months ago

#30 (comment)

因为原始中文数据集只有两个类别。

非常感谢！

AsuSociety commented 4 months ago

For the datasets, each row contains a token or word, an ATE tag and a polarity tag. If a token or word does not belong to an aspect, the polarity is -1, else: 0 for negative， 1 for neutral， 2 for positive.

However, each sample only label polarity for one aspect, other's polarity is -1 the same as non aspect words.

e.g.

I O -1 charge O -1 it O -1 at O -1 night O -1 and O -1 skip O -1 taking O -1 the O -1 cord B-ASP 1 with O -1 me O -1 because O -1 of O -1 the O -1 good O -1 battery B-ASP -1 life I-ASP -1 . O -1

Hi, I'm trying to prepare a data set for a project I'm doing and I couldn't understand the reference to the ATE tag, when does it get 0/1/2? Compared to the example you gave here, my data set is divided into complete sentences that return positive/negative (in polarity).

Thanks in advance.