synthetic-healthcare-admissions

# Synthetic Healthcare Admissions Dataset

### A fully synthetic **healthcare dataset** for building AI solutions in healthcare, developed using Syncora.ai.

---

## ✅ What's in This Repo?

- ✅ **Healthcare Dataset (CSV)** → [Download Here](https://huggingface.co/datasets/syncora/synthetic-healthcare-admissions/blob/main/Healthcare_Syncora_Synthetic%201.csv)
- ✅ **Example Jupyter Notebook** → [Open Notebook](https://huggingface.co/datasets/syncora/synthetic-healthcare-admissions/blob/main/Healthcare_Syncora_Synthetic_1%20(1).ipynb)
- ✅ **Use cases**

---

## 📘 About This Dataset

This **synthetic healthcare dataset** simulates hospital admission records including demographics, billing, medications, and lab results.
It is **100% synthetic**, ensuring privacy and regulatory compliance for developers, healthcare institutes, and those training LLM models.

**Why use this dataset?**
- Explore **predictive modeling in healthcare**
- Build **dataset for LLM training** for clinical conversations
- Safely **generate synthetic data** without exposing real patient info

---

## 🔍 Dataset Snapshot

| Column | Description |
|--------------------|----------------------------------------------------|
| **Age** | Patient age in years |
| **Gender** | 0 = Female, 1 = Male |
| **Blood Type** | Encoded blood group category (0–7) |
| **Medical Condition** | Encoded diagnosis category |
| **Billing Amount** | Hospital billing in USD |
| **Admission Type** | 0 = Emergency, 1 = Urgent, 2 = Elective |
| **Medication** | Encoded medication type |
| **Test Results** | Encoded lab test result category |

**Example row:**
`80, 1, 7, 0, 37303.07, 0, 0, 0`

---

## ✅ Use Cases

This **healthcare dataset** is ideal for:

- 🏥 **Predictive Healthcare Analytics** – Predict billing amount, admission type, or risk scores
- 💊 **Medication Optimization Models** – Analyze treatment outcomes
- 🗣 **Healthcare Chatbots** – Train conversational LLMs on realistic medical workflows
- 📊 **Cost Forecasting** – Estimate hospital expenses
- 🧠 **Dataset for LLM Training** – Fine-tune models for clinical Q&A or triage

---

## 🚀 Generate Your Own Synthetic Data

Need custom scenarios? Use our tool to generate synthetic data tailored to your requirements:
👉 [**Generate your own Synthetic Data Now**](https://huggingface.co/spaces/syncora/synthetic-generation)

---

## ⚡ Quick Start

```python
from datasets import load_dataset

dataset = load_dataset("syncora/synthetic-healthcare-admissions")
df = dataset["train"].to_pandas()
print(df.head())

# 合成医疗入院数据集

### 完全合成的医疗数据集（healthcare dataset），依托Syncora.ai开发，用于构建医疗领域人工智能解决方案。

---

## ✅ 仓库内容概览

- ✅ **医疗数据集（CSV格式）** → [点击下载](https://huggingface.co/datasets/syncora/synthetic-healthcare-admissions/blob/main/Healthcare_Syncora_Synthetic%201.csv)
- ✅ **示例Jupyter Notebook** → [打开 Notebook](https://huggingface.co/datasets/syncora/synthetic-healthcare-admissions/blob/main/Healthcare_Syncora_Synthetic_1%20(1).ipynb)
- ✅ **应用场景**

---

## 📘 数据集简介

本**合成医疗数据集（synthetic healthcare dataset）**可模拟医院入院记录，涵盖人口统计学信息、账单明细、用药情况与实验室检测结果。
该数据集为100%合成生成，可保障开发者、医疗机构以及训练大语言模型（Large Language Model，LLM）的人员的数据隐私与合规性。

**为何选择本数据集？**
- 开展医疗领域预测建模研究
- 构建用于临床对话场景的大语言模型训练数据集
- 在不泄露真实患者信息的前提下，安全生成合成数据

---

## 🔍 数据集快照

| 列名 | 描述说明 |
|--------------------|----------------------------------------------------|
| **年龄（Age）** | 患者年龄（单位：岁） |
| **性别（Gender）** | 0 = 女性，1 = 男性 |
| **血型（Blood Type）** | 编码后的血型分类（取值范围0~7） |
| **病症类型（Medical Condition）** | 编码后的诊断分类 |
| **账单金额（Billing Amount）** | 医院账单金额（单位：美元） |
| **入院类型（Admission Type）** | 0 = 急诊，1 = 加急，2 = 择期 |
| **用药类型（Medication）** | 编码后的用药类别 |
| **检测结果（Test Results）** | 编码后的实验室检测结果分类 |

**示例数据行：**
`80, 1, 7, 0, 37303.07, 0, 0, 0`

---

## ✅ 典型应用场景

本医疗数据集适用于以下场景：

- 🏥 **预测性医疗分析** – 预测账单金额、入院类型或风险评分
- 💊 **用药优化模型** – 分析治疗效果
- 🗣 **医疗聊天机器人** – 基于真实医疗流程训练对话式大语言模型
- 📊 **成本预测** – 估算医院运营开支
- 🧠 **大语言模型训练数据集** – 微调模型以支持临床问答或分诊服务

---

## 🚀 生成自定义合成数据

需要定制化场景？可使用我们的工具生成符合您需求的合成数据：
👉 [**立即生成专属合成数据**](https://huggingface.co/spaces/syncora/synthetic-generation)

---

## ⚡ 快速入门

python
from datasets import load_dataset

dataset = load_dataset("syncora/synthetic-healthcare-admissions")
df = dataset["train"].to_pandas()
print(df.head())

搜集汇总

背景与挑战

背景概述

该数据集是一个完全合成的医疗数据集，模拟医院入院记录，涵盖年龄、性别、医疗状况、账单金额等字段，旨在支持医疗AI开发、预测分析和LLM训练，同时确保数据隐私和合规性。它适用于医疗聊天机器人、成本预测和药物优化等多种应用场景。