A two-stage Q-learning algorithm jointly chooses treatments and which covariates to collect, maximizing expected outcome minus assessment and treatment costs.
We specify the working model for ∆ 1c(sl1, j1) 43 as: ∆1c∗(sl1, j1) = s⊤ l1δ∗ j1, δ∗ j1 = arg min δ E{(Y 1c∗ j1 − Y 1c∗ jf 1 − S⊤ l1δ)2}
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ME 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Balancing utility and cost in dynamic treatment regimes
A two-stage Q-learning algorithm jointly chooses treatments and which covariates to collect, maximizing expected outcome minus assessment and treatment costs.