PU-NA-BERT数据集

数据集获取（Figshare）

• 源站标题：PU-NA-BERT
• DOI：10.6084/m9.figshare.21667718.v1
• 发布日期：2022-12-02
• 文件数：2 个
• 文件总体积：720.29 KB
• 源站页面：https://figshare.com/articles/dataset/PU-NA-BERT/21667718

命令行下载

curl -L -o "data.rar" "https://ndownloader.figshare.com/files/38411066"
curl -L -o "procedure.rar" "https://ndownloader.figshare.com/files/38411093"

文件清单

• data.rar（563.95 KB）
• procedure.rar（156.34 KB）

文件总体积 720.29 KB（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

We used Electronic medical records from public use datasets provided by Yidu Cloud to obtain the distant dataset by dista。nt supervision. For the obtained distant dataset, to enhance the reliability of the dataset, we used PU learning for denoising to reduce the negative impact of incorrectly labelled negative samples or unlabeled samples on the model. Finally, cancer information extraction was performed using negative sampling and pre-trained modelsWe used Electronic medical records from public use datasets provided by Yidu Cloud to obtain the distant dataset by distant supervision. For the obtained distant dataset, to enhance the reliability of the dataset, we used PU learning for denoising to reduce the negative impact of incorrectly labelled negative samples or unlabeled samples on the model. Finally, cancer information extraction was performed using negative sampling and pre-trained models

许可

CC BY 4.0

引用

学术使用请引用 DOI 10.6084/m9.figshare.21667718.v1（source: https://figshare.com/articles/dataset/PU-NA-BERT/21667718）。