ictwjw / Ward2ICU

Ward2ICU: A Vital Signs Dataset of Inpatients from the General Ward

---

Ward2ICU

• Description
• Models
  • 1D Conditional CNN GAN
  • Recursive GAN (RGAN)
  • Recursive Conditional GAN (RCGAN)
  • RNN Classifier
  • 1D-CNN Classifier
• Citation

Description

Ward2ICU: A Vital Signs Dataset of Inpatients from the General Ward

Models

1D Conditional CNN GAN

Recursive GAN (RGAN)

Recursive GAN (Generator) implementation with RNN cells.

Recursive Conditional GAN (RCGAN)

Recursive Conditional GAN (Generator) implementation with RNN cells

RNN Classifier

A simple RNN for classification tasks. It consists of a recurrent layer (Elman RNN, LSTM or GRU) followed by 2 fully connected. The first shares parameters across the time domain (i.e. second tensor dimension), while the second collapses the time-domain to a single point with a Sigmoid activation.

1D-CNN Classifier

Single-dimension convolutional network for classification. Consists of a sequence of Conv1d followed by MaxPool1d and Linear with a Sigmoid output.

Citation

@article{severo2019ward2icu,
  title={Ward2ICU: A Vital Signs Dataset of Inpatients from the General Ward},
  author={Severo, Daniel and Amaro, Fl{\'a}vio and Hruschka Jr, Estevam R and Costa, Andr{\'e} Soares de Moura},
  journal={arXiv preprint arXiv:1910.00752},
  year={2019}
}