en-fpb-ko / README.md
kbshin's picture
Update README.md
543e8cc verified
---
dataset_info:
features:
- name: conversation_id
dtype: string
- name: conversations
list:
- name: from
dtype: string
- name: value
dtype: string
splits:
- name: test
num_bytes: 333061
num_examples: 944
download_size: 109275
dataset_size: 333061
configs:
- config_name: default
data_files:
- split: test
path: data/test-*
license: mit
language:
- ko
---
# en-fpb-ko
### 데이터 설명
- `en-fpb-ko` 데이터는 금융 뉴스로부터의 문장을 '긍정', '중립', '부정' 중 하나로 분류하는 감성 분류 (sentiment analysis) 데이터셋입니다.
입력값으로는 text만 주어집니다.
- 한국어 데이터를 생성하기 위해, 먼저 사내 언어 번역 모델인 Allganize Translator를 활용하여 [ChanceFocus/en-fpb](https://huggingface.co/datasets/ChanceFocus/en-fpb)의 test set을 한국어로 번역하였습니다.
오역된 데이터를 직접 제거하였고, 그 결과 944개의 평가 데이터가 생성되었습니다.
### 데이터 출처
- [ChanceFocus/en-fpb](https://huggingface.co/datasets/ChanceFocus/en-fpb)
### 데이터 예시
```
{
'conversation_id': 'fpb3876',
'conversations': array([
{
'from': 'human',
'value': '''금융 뉴스 기사에서 발췌한 이 문장의 정서를 분석하세요. 부정, 긍정 또는 중립 중 하나로 답을 제시하세요.
텍스트: 회사 간의 오랜 협력 관계를 확대하는 새로운 계약에는 Larox에서 Etteplan으로 특정 엔지니어링 및 문서 기능의 이전이 포함됩니다.
정답:'''
},
{
'from': 'gpt',
'value': '긍정'
}
], dtype=object)
}
```