Skip to content
/ BRITS Public

Code of NIPS18 Paper: BRITS: Bidirectional Recurrent Imputation for Time Series

License

Notifications You must be signed in to change notification settings

caow13/BRITS

Folders and files

NameName
Last commit message
Last commit date

Latest commit

fc0a3a4 · Dec 15, 2018

History

8 Commits
Oct 11, 2018
Nov 28, 2018
Oct 11, 2018
Dec 12, 2018
Dec 15, 2018
Oct 11, 2018
Oct 11, 2018
Oct 11, 2018
Oct 11, 2018
Oct 11, 2018
Oct 11, 2018

Repository files navigation

README

It is a pytorch implemention of paper "BRITS: Bidirectional Recurrent Imputation for Time Series, Wei Cao, Dong Wang, Jian Li, Hao Zhou, Lei Li Yitan Li. (NerIPS 2018)". The paper can be found here. http://papers.nips.cc/paper/7911-brits-bidirectional-recurrent-imputation-for-time-series

To train the BRIST model, first please unzip the PhysioNet data into raw folder, including the label file Outcomes-a.txt.

To run the model:

  • make a empty folder named json, and run inpute_process.py.
  • run different models:
    • e.g., RITS_I: python main.py --model rits_i --epochs 1000 --batch_size 64 --impute_weight 0.3 --label_weight 1.0 --hid_size 108
    • for most cases, using impute_weight=0.3 and label_weight=1.0 lead to a good performance. Also adjust hid_size to control the number of parameters

About

Code of NIPS18 Paper: BRITS: Bidirectional Recurrent Imputation for Time Series

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages