Dynamic Treatment Regimes

---

Topics and Objectives (Apr 1 – Apr 5)

• Data generating process for dynamic treatment regimes
• Dynamic programming and batch Q-learning
• Generalization Bound for batch Q-learning
• Adapting OWL in sequential decisions
• Using the DynTxRegime package to implement these methods

---

Lecture Notes

• Batch Q-learning for Dynamic Treatment Regime
• R Example: Batch Q-learning for Finite Stage
• Outcome Weighted Learning for Dynamic Decisions
• R Example: Backward Outcome Weighted Learning

---

Homework

• See Homework 4 for details
• Homework 4 covers week 12 (this week) and 13 (MDP)

---