YAIB
YAIB copied to clipboard
rvandewater
🧪Yet Another ICU Benchmark: a holistic framework for the standardization of clinical prediction model experiments. Provide custom datasets, cohorts, prediction tasks, endpoints, preprocessing, and mo...
We want users to be able to specify a type of balanced sampling. So for specifying test and train set, we would have the same ratio of positive and negative...
Using the `demo_data`, write tests to make sure both `train` and `evaluate` work correctly. Trigger these tests on commits to PRs.
See: _Medical Data Wrangling With Sequential Variational Autoencoders,' in IEEE Journal of Biomedical and Health Informatics, vol. 26, no. 6, pp. 2737-2745, June 2022, doi: 10.1109/JBHI.2021.312383,_ GitHub repo: https://github.com/dbarrejon/Shi-VAE/blob/e03a1a727df0489b67d23cd53664fadadc5b3032/src/lib/loss.py#L18
Use parallelized library Dask in place of numpy and pandas. https://www.dask.org/
Calibrate models with isotonic regressor after training them
Suggested by Hendrik to improve the software engineering soundness
Metadata
Owner
Metadata
🧪Yet Another ICU Benchmark: a holistic framework for the standardization of clinical prediction model experiments. Provide custom datasets, cohorts, prediction tasks, endpoints, preprocessing, and mo...