Bangla Healthcare Severity Dataset

Published: 26 June 2026| Version 1 | DOI: 10.17632/37z8kgx79m.1
Contributors:
,
,
,

Description

The Bangla Healthcare Severity Dataset is a manually annotated Bangla healthcare text corpus developed for healthcare Natural Language Processing (NLP) research. The dataset contains 5,263 real-world Bangla healthcare-related queries collected from publicly accessible Facebook and YouTube discussions. Each sample has been manually annotated into one of four healthcare severity classes: Emergency, Urgent, Routine, and General Query. The dataset is intended to support research in Bangla NLP, healthcare artificial intelligence, medical text classification, clinical decision support systems, and transformer-based language models. Dataset Summary • Total Samples: 5,263 • Language: Bangla • Classes: 4 • Annotation: Manual Columns • Text • Categories • Action Needed Applications • Bangla Natural Language Processing • Healthcare NLP • Medical Text Classification • Clinical Severity Classification • Healthcare Chatbots • Clinical Decision Support Systems • Transformer-based Deep Learning • Large Language Models (LLMs) Data Collection The data were collected from publicly accessible healthcare-related discussions on Facebook and YouTube. Personally identifiable information (PII) was removed during preprocessing. Every sample was manually annotated and quality-checked before inclusion in the dataset. License Creative Commons Attribution 4.0 International (CC BY 4.0) Citation Pulock Deb Roy, Mst. Chameli Begum, Dolour Husain Sagor, and Badar Hossain. Bangla Healthcare Severity Dataset. Mendeley Data. Version 1. 2026.

Files

Steps to reproduce

1. Open the Bangla-Healthcare_Dataset.csv or Bangla-Healthcare_Dataset.xlsx file. 2. The dataset contains three columns: - Text - Categories - Action Needed 3. Load the dataset using Python, R, MATLAB, or any spreadsheet software. 4. Split the data into training, validation, and testing sets as required. 5. Use the "Categories" column as the target label for supervised machine learning or deep learning models. 6. The dataset can be used for healthcare severity classification, healthcare NLP, medical text classification, and Bangla language processing research.

Institutions

Categories

Computer Science, Artificial Intelligence, Health Informatics, Natural Language Processing

Licence