metadata
language:
- en
license: apache-2.0
task_categories:
- text-generation
- question-answering
tags:
- health
- english
- kicaulah-ai
- natural-language
- medical
- health
- pediatric
- first-aid
- mental-health
- clinical-qa
- wellness
pretty_name: Kicaulah AI - Dataset Health (Medical QA & Triage)
size_categories:
- 1K<n<10K
Kicaulah AI — Dataset Health (Medical QA & Triage)
📖 Description
A high-quality Medical QA and Health Education dataset in natural conversational English. Covers 10 distinct clinical pillars: Pediatrics & Infant Care, Gastric & GERD, Chronic Metabolic Disease (Hypertension, Diabetes, Gout, Cholesterol), Emergency First Aid, Dermatology & Skin Health, Women's Health, Mental Wellness & Sleep Hygiene, Over-The-Counter Pharmacology, Oral & Dental Care, and Respiratory & Infectious Disease. Designed under the Anti-AI-Speak standard with human empathy and zero robotic clichés.
This dataset is strictly curated under the Anti-AI-Speak Standard:
- Built using authentic, everyday conversational language.
- Free from generic robotic clichés such as "As an AI language model..." or "That's a great question!".
- Prioritizes genuine empathy, real-world context, practical analogies, and natural variations in tone (casual, formal, emotional, concise, and in-depth).
🎯 Use Cases
- Fine-tuning Large Language Models (LLMs) for natural, empathetic medical assistant applications.
- Training health triage bots to deliver concise, reassuring, and red-flag-conscious medical guidance.
- Benchmarking conversational healthcare responses against everyday, realistic patient inquiries.
📊 Dataset Structure
| Column | Type | Description | Example |
|---|---|---|---|
| instruction | string | User inquiry or realistic scenario | "My 2-year-old has had a 102F fever for 2 days, what should I do?" |
| response | string | Natural, human, solution-oriented response | "Take a deep breath first. A 102 fever on day 2 is scary..." |
| category | string | Specific domain subcategory | "health" |
📈 Statistics
- Total Examples: 1,100
- Train (80%): 880
- Validation (10%): 110
- Test (10%): 110
🚀 How to Use
from datasets import load_dataset
dataset = load_dataset("Kicaulah/dataset-health")
print(dataset["train"][0])
📑 Citation
@misc{kicaulah_health,
title={Kicaulah AI - Dataset Health (Medical QA & Triage)},
author={Kicaulah AI Team},
year={2025},
publisher={Hugging Face},
howpublished={\url{https://huggingface.co/datasets/Kicaulah/dataset-health}}
}