language:
- en
license: cc-by-sa-4.0
size_categories:
- n<1K
tags:
- medical
dataset_info:
features:
- name: sentences
sequence: string
- name: labels
sequence: string
splits:
- name: test
num_bytes: 304495
num_examples: 1
download_size: 126661
dataset_size: 304495
configs:
- config_name: default
data_files:
- split: test
path: data/test-*
CLUSTREC-COVID: A Topical Clustering Benchmark for COVID-19 Scientific Research
Dataset Summary
CLUSTREC-COVID-S2S is a modified version of the TREC-COVID dataset, transformed into a topical clustering benchmark. The dataset consists of titles and abstracts from scientific papers about COVID-19 research, covering a diverse range of research topics. Each document in the dataset is assigned to a specific subtopic, making it ideal for use in document clustering and topic modeling tasks.
The dataset is useful for researchers aiming to evaluate clustering algorithms and techniques for automatic organization of scientific literature. It can also be used for exploring information retrieval systems that aim to group documents by subtopic or related research areas.
The source of this dataset is the TREC-COVID retrieval dataset, which has been adapted for clustering and organization tasks.
Dataset Structure
Each document in the dataset includes the following fields:
- labels (string): The specific subtopic to which the document has been assigned. (e.g., "coronavirus response to weather changes").
- sentences (string): The title of the scientific paper.
- doc_id (string): A unique document identifier.
Example Entry
{
"labels": "coronavirus response to weather changes",
"sentences": "Weather variables impact on COVID-19 incidence",
"doc_id": "hadnxjeo",
}
Citation Information
Cite as:
@article{katz2024knowledge,
title={Knowledge Navigator: LLM-guided Browsing Framework for Exploratory Search in Scientific Literature},
author={Katz, Uri and Levy, Mosh and Goldberg, Yoav},
journal={arXiv preprint arXiv:2408.15836},
year={2024}
}