Skip to content

Dataset · owaiskha9654

PubMed research articles for multi-label classification

owaiskha9654/PubMed_MultiLabel_Text_Classification_Dataset_MeSH

Around 50,000 English-language medical articles from PubMed with manual expert annotation using the MeSH term hierarchy.

What's inside

The dataset contains around 50,000 scientific publications in English from the PubMed repository, presented in text and tabular formats. Initially, biomedical experts manually assigned 10 to 15 thematic MeSH labels to each article. To reduce data sparsity and an overly large class space, the authors processed the labels and mapped them to root categories.

What you can do with it

  • Training models for multi-label classification of scientific and biomedical texts
  • Building automated indexing and thematic search systems for medical articles
Tasks
Text classification
Languages
English
Size
10K–100K rows
Downloads this month
1.2K
Likes
41

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("owaiskha9654/PubMed_MultiLabel_Text_Classification_Dataset_MeSH")