基于FAISS的临床糖尿病数据集向量嵌入

数据集获取（Figshare）

• 源站标题：FAISS-Based Vector Embeddings of Clinical Diabetes Dataset for LLM-Driven Analytical Systems
• DOI：10.6084/m9.figshare.31832887.v1
• 发布日期：2026-03-23
• 文件数：1 个
• 文件总体积：14.79 MB
• 源站页面：https://figshare.com/articles/dataset/FAISS-Based_Vector_Embeddings_of_Clinical_Diabetes_Dataset_for_LLM-Driven_Analytical_Systems/31832887

命令行下载

curl -L -o "faiss_diabetes_index.zip" "https://ndownloader.figshare.com/files/63010300"

文件清单

• faiss_diabetes_index.zip（14.79 MB）

文件总体积 14.79 MB（源站 API 实测）。

数据加载示例（图像类）

from PIL import Image
import glob, os

files = (glob.glob(os.path.join(path, "**", "*.png"), recursive=True)
       + glob.glob(os.path.join(path, "**", "*.jpg"), recursive=True)
       + glob.glob(os.path.join(path, "**", "*.tif"), recursive=True))
print("图像文件数:", len(files))
img = Image.open(files[0]); print("尺寸/模式:", img.size, img.mode)

# torchvision Dataset 方式:
# from torchvision import datasets
# ds = datasets.ImageFolder(path)  # 要求 子目录=类别

目录组织与标注格式以源站说明和下载后实际文件为准。

数据说明

This dataset provides a FAISS-based vector database constructed from a structured clinical diabetes dataset to support advanced Artificial Intelligence (AI) and Large Language Model (LLM)-driven analytical workflows. The original tabular patient records, including demographic and clinical attributes such as age, gender, body mass index (BMI), HbA1c levels, blood glucose levels, and comorbid conditions, were transformed into structured textual representations to enable semantic processing.
These textual records were embedded using the all-MiniLM-L6-v2 sentence transformer model, generating dense vector representations suitable for similarity search and retrieval tasks. The embeddings were indexed using the FAISS (Facebook AI Similarity Search) library, enabling efficient large-scale vector retrieval for Retrieval-Augmented Generation (RAG) pipelines and agent-based reasoning systems.
The dataset is specifically designed to support reproducible research in:

• Retrieval-Augmented Generation (RAG)
• Agentic AI systems with tool execution
• Clinical decision support simulations
• Semantic search over structured healthcare data
• Performance evaluation of LLM-driven analytical agents
  Additionally, the dataset supports experimental analysis of system-level characteristics such as latency, scalability, and query-response behavior when integrated with LLM frameworks. It is suitable for benchmarking, prototyping, and evaluating AI systems that operate on structured clinical data through semantic abstraction.
  This resource enables interdisciplinary research at the intersection of artificial intelligence, data science, and healthcare analytics, facilitating the development of intelligent systems for clinical insight generation and data-driven decision-making.

许可

CC BY 4.0

引用

学术使用请引用 DOI 10.6084/m9.figshare.31832887.v1（source: https://figshare.com/articles/dataset/FAISS-Based_Vector_Embeddings_of_Clinical_Diabetes_Dataset_for_LLM-Driven_Analytical_Systems/31832887）。