NIILC-ECQA2015
收藏资源简介:
该数据集是一个用于支持问答系统开发的数据集,包含问题及其答案,以及诸如问题类型、获取答案的线索、包含答案的维基百科页面和JWO(日本维基百科本体)的SPARQL查询等元信息。数据集包含两个配置:default和v1.1。default配置包含800个开发样本和200个测试样本,而v1.1配置包含198个测试样本。数据集的主要特征包括id、question、answers以及一系列其他字段,如B1、B2等,这些字段可能包含与问题和答案相关的额外信息。数据集的语言为日语,许可证为CC BY-SA 4.0。
This dataset is developed to support the construction of question answering (QA) systems. It contains questions and their corresponding answers, along with metadata including question type, clues for acquiring answers, Wikipedia pages that host the target answers, and SPARQL queries based on the Japanese Wikipedia Ontology (JWO). The dataset provides two configurations: default and v1.1. The default configuration includes 800 development samples and 200 test samples, while the v1.1 configuration contains 198 test samples. Core fields of the dataset include id, question, answers, as well as a series of supplementary fields such as B1, B2, etc., which may store additional information related to the questions and answers. The dataset is in Japanese, and it is licensed under CC BY-SA 4.0.
NIILC Question Answering Dataset
概述
该数据集用于支持问答系统的发展,包含问题及其答案,以及相关的元信息,如问题类型、获取答案的线索、答案所在的维基百科页面和JWO(日本维基百科本体)的SPARQL查询。
数据字段
id(str): 唯一ID。text(str): 问题文本。answers(list[str]): 问题的所有可接受答案。B1至B7(str): 字符串类型。C_1,C_2(str): 字符串类型。D_2至D_4(str): 字符串类型。E_1至E_8(str): 字符串类型。F_1,F_2,F_memo(str): 字符串类型。C_memo,E_memo,A_memo,D_memo,A_3(str): 字符串类型。
数据集配置
默认配置
- 配置名称:
default - 数据文件:
dev:data/dev-*test:data/test-*
- 分割:
dev: 800个样本,491997字节test: 200个样本,123513字节
- 下载大小: 336871字节
- 数据集大小: 615510字节
v1.1配置
- 配置名称:
v1.1 - 数据文件:
test:v1.1/test-*
- 分割:
test: 198个样本,123340字节
- 下载大小: 82101字节
- 数据集大小: 123340字节
许可信息
Creative Commons Attribution Share Alike 4.0 International
引用信息
@inproceedings{sekine2003question, title={Development of a question answering system focused on an encyclopedia}, author={Sekine, Satoshi}, booktitle={9th Annual Meeting of the Association for Natural Language Processing}, year={2003}, language={Japanese} }




