irds/nyt_wksup_train

`nyt/wksup/train`数据集由ir-datasets包提供，主要用于文本检索任务。该数据集包含查询（queries）和相关性评估（qrels）两部分数据，查询数量为1,863,657条，相关性评估数量也为1,863,657条。文档部分需要从`irds/nyt`数据集中获取。数据集的使用可以通过Hugging Face的`load_dataset`函数进行加载和访问。

The `nyt/wksup/train` dataset, provided by the ir-datasets package, is primarily utilized for text retrieval tasks. It contains two components: queries and relevance judgments (qrels). The number of queries is 1,863,657, and the count of relevance judgments is also 1,863,657. The document corpus needs to be obtained from the `irds/nyt` dataset. This dataset can be loaded and accessed via Hugging Face's `load_dataset` function.

原始信息汇总

数据集概述

数据集名称

nyt/wksup/train

数据来源

• 源数据集: irds/nyt

任务类别

• 文本检索

数据内容

• queries: 查询（主题），数量为1,863,657
• qrels: 相关性评估，数量为1,863,657
• docs: 文档数据，需从irds/nyt获取

使用示例

python from datasets import load_dataset

queries = load_dataset(irds/nyt_wksup_train, queries) for record in queries: record # {query_id: ..., text: ...}

qrels = load_dataset(irds/nyt_wksup_train, qrels) for record in qrels: record # {query_id: ..., doc_id: ..., relevance: ...}

引用信息

@inproceedings{MacAvaney2019Wksup, author = {MacAvaney, Sean and Yates, Andrew and Hui, Kai and Frieder, Ophir}, title = {Content-Based Weak Supervision for Ad-Hoc Re-Ranking}, booktitle = {SIGIR}, year = {2019} } @article{Sandhaus2008Nyt, title={The new york times annotated corpus}, author={Sandhaus, Evan}, journal={Linguistic Data Consortium, Philadelphia}, volume={6}, number={12}, pages={e26752}, year={2008} }