A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf
.
Clairvoyance: A Pipeline Toolkit for Medical Time Series
2021
International Conference on Learning Representations
Time-series learning is the bread and butter of data-driven clinical decision support, and the recent explosion in ML research has demonstrated great potential in various healthcare settings. At the same time, medical time-series problems in the wild are challenging due to their highly composite nature: They entail design choices and interactions among components that preprocess data, impute missing values, select features, issue predictions, estimate uncertainty, and interpret models. Despite
dblp:conf/iclr/JarrettYBQES21
fatcat:ltp7ijb24rhqnccpby74cooflq