Datasets:

IMISLab
/

GreekPubMed4000

license: apache-2.0
task_categories:
  - text-classification
language:
  - el
tags:
  - greek
  - nlp
  - clustering
  - classification
  - medical
pretty_name: GreekPubMed4000
size_categories:
  - 1K<n<10K
configs:
  - config_name: default
    data_files:
      - split: all_rows
        path: GreekPubMed4000.csv

GreekPubMed4000

We introduce GreekPubMed40000, a clustering and classification dataset that comprises 4,000 medical documents translated into Greek from PubMed4000. For information about dataset creation, limitations etc. see the upcoming conference paper - TBA.

Supported Tasks

This dataset supports Text Clustering and Classification.

Language

All dataset samples are written in Greek.

Dataset Structure

The dataset is structured as a .csv file of 4000 rows.
The following data fields are provided:

id: (int) The row id.
text: (str) The medical text.
label: (str): The classification label (4 different values).

Example code

from datasets import load_dataset

# Load the dataset.
dataset = load_dataset('IMISLab/GreekPubMed4000', split = 'all_rows')
print(dataset[0])

Contact

If you have any questions/feedback about the dataset please e-mail one of the following authors:

giarelis@ceid.upatras.gr
karacap@upatras.gr

Citation

TBA