Simulations of three weakly-supervised phenotyping methods show variable performance in rare-outcome EHR settings, with sureLDA often strong but all sensitive to tuning and silver-label strength.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ME 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Performance of weakly-supervised electronic health record-based phenotyping methods in rare-outcome settings
Simulations of three weakly-supervised phenotyping methods show variable performance in rare-outcome EHR settings, with sureLDA often strong but all sensitive to tuning and silver-label strength.