Preview

You are previewing a new record that has not yet been published.

Published February 21, 2025 | Version 1.0.0
Dataset Open

LongEval 2025 Web Retrieval Collection

Description

The collection comprises queries and documents sourced from the Qwant search engine. The user-issued queries focus on selected trending topics. Documents were chosen based on these queries using Qwant's click model, with additional randomly selected documents from Qwant's index. Data collection occurred from June 2022 to February 2023, resulting in 9,000 training queries and 18 million associated documents. Relevance assessments, derived from the Qwant click model, are provided for these queries. The collection includes the original French version of the webpages and queries. This dataset serves as the official training collection for the 2025 LongEval Information Retrieval Lab, organized at CLEF. Full info at https://clef-longeval.github.io/

This data set contains both the training and the test data.

The relevance assessments for the test topics will be added after the evaluation cycle has ended.

Files

README.pdf

Files (124.2 GiB)

Name Size
md5:8de0e575fb7aa735e98d8d8975d6bc2d
36.9 GiB Preview Download
md5:a84d2b2b48190298b5337a67f9ece026
36.2 GiB Preview Download
md5:6e87d6063aa1d6db402fd5099163bfd4
51.1 GiB Preview Download
md5:66853c99c51d641cfcf2da9a6c277740
285.0 KiB Preview Download

Additional details

Related works

Continues
Dataset: 10.48436/y60e9-k9b51 (DOI)
Dataset: 10.48436/xr350-79683 (DOI)

Funding

Kodicare I 4471-N
FWF Austrian Science Fund