- 설명 :
학술 논문의 인용 의도를 분류하기 위한 데이터셋입니다. 각 Json 개체의 기본 인용 의도 레이블은 레이블 키로 지정되고 인용 컨텍스트는 컨텍스트 키로 지정됩니다. 예시:
{
'string': 'In chacma baboons, male-infant relationships can be linked to both
formation of friendships and paternity success [30,31].'
'sectionName': 'Introduction',
'label': 'background',
'citingPaperId': '7a6b2d4b405439',
'citedPaperId': '9d1abadc55b5e0',
...
}
Semantic Scholar API( https://api.semanticscholar.org/ )와 함께 제공된 논문 ID를 사용하여 논문에 대한 전체 정보를 얻을 수 있습니다.
레이블은 메서드, 배경, 결과입니다.
추가 문서 : 코드가 있는 논문에서 탐색
소스 코드 :
tfds.datasets.scicite.Builder버전 :
-
1.0.0(기본값): 릴리스 정보가 없습니다.
-
다운로드 크기 :
22.12 MiB데이터 세트 크기 :
7.26 MiB자동 캐시 ( 문서 ): 예
분할 :
| 나뉘다 | 예 |
|---|---|
'test' | 1,859 |
'train' | 8,194 |
'validation' | 916 |
- 기능 구조 :
FeaturesDict({
'citeEnd': int64,
'citeStart': int64,
'citedPaperId': Text(shape=(), dtype=string),
'citingPaperId': Text(shape=(), dtype=string),
'excerpt_index': int32,
'id': Text(shape=(), dtype=string),
'isKeyCitation': bool,
'label': ClassLabel(shape=(), dtype=int64, num_classes=3),
'label2': ClassLabel(shape=(), dtype=int64, num_classes=4),
'label2_confidence': float32,
'label_confidence': float32,
'sectionName': Text(shape=(), dtype=string),
'source': ClassLabel(shape=(), dtype=int64, num_classes=7),
'string': Text(shape=(), dtype=string),
})
- 기능 문서 :
| 특징 | 수업 | 모양 | D타입 | 설명 |
|---|---|---|---|---|
| 풍모Dict | ||||
| 인용종료 | 텐서 | int64 | ||
| 인용시작 | 텐서 | int64 | ||
| 인용용지 ID | 텍스트 | 끈 | ||
| 인용용지 ID | 텍스트 | 끈 | ||
| 발췌_색인 | 텐서 | int32 | ||
| ID | 텍스트 | 끈 | ||
| isKeyCitation | 텐서 | 부울 | ||
| 상표 | 클래스 레이블 | int64 | ||
| 라벨2 | 클래스 레이블 | int64 | ||
| label2_confidence | 텐서 | float32 | ||
| label_confidence | 텐서 | float32 | ||
| 섹션 이름 | 텍스트 | 끈 | ||
| 원천 | 클래스 레이블 | int64 | ||
| 끈 | 텍스트 | 끈 |
감독 키 (
as_supervised문서 참조):('string', 'label')그림 ( tfds.show_examples ): 지원되지 않습니다.
예 ( tfds.as_dataframe ):
- 인용 :
@InProceedings{Cohan2019Structural,
author={Arman Cohan and Waleed Ammar and Madeleine Van Zuylen and Field Cady},
title={Structural Scaffolds for Citation Intent Classification in Scientific Publications},
booktitle="NAACL",
year="2019"
}