Targeted Learning in Data Science

Part I: Introductory Chapters

1. The Statistical Estimation Problem in Complex Longitudinal Data

Data Science and Statistical EstimationRoadmap for Causal Effect EstimationRole of Targeted Learning in Data ScienceObserved DataCaussal Model and Causal target QuantityStatistical ModelStatistical Target ParameterStatistical Estimation Problem















2. Longitudinal Causal Models

Structural Causal ModelsCausal Graphs / DAGsNonparametric Structural Equation Models





3. Super Learner for Longitudinal Problems

Ensemble LearningSequential Regression



4. Longitudinal Targeted Maximum Likelihood Estimation (LTMLE)

Step-by-Step Demonstration of LTMLE



5. Understanding LTMLE

Statistical PropertiesTheoretical Background



6. Why LTMLE?

Landscape of Other EstimatorsComparison of Statistical Properties





Part II: Additional Core Topics

7. One-Step TMLE

General FrameworkTheoretical Results



8. One-Step TMLE for the Effect Among the Treated

Demonstration for Effect Among the TreatedSimulation Studies



9. Online Targeted Learning

Batched Streaming DataOnline and One-Step EstimatorTheoretical Considerations





10. Networks

General Statistical FrameworkCausal Model for Network DataCounterfactual Mean Under Stochastic Intervention on the NetworkDevelopment of TMLE for NetworksInference







11. Application to Networks

Differing Network StructuresRealistic Network Examples (e.g., effect of vaccination)R Package Implementation of TMLE





12. Targeted Estimation of the Nuisance Parameter

Asymptotic LinearityIPWTMLE





13. Sensitivity AnalysesGeneral Nonparametric Approach to Sensitivity AnalysisMeasurement ErrorUnmeasured ConfoundingInformative Missingness of the OutcomeFDA Meta-Analysis











Part III: Randomized Trials

14. Community Randomized Trials for Small Samples

Introduction of SEARCH Community Randomized TrialAdaptive Pair MatchingData-Adaptive Selection of Covariates for Small SamplesTMLE Using Super Learning for Small SamplesInference









15. Sample Average Treatment Effect in a CRT

Introduction of the ParameterEffect for the Observed CommunitiesInference





16. Application to Clinical Trial Survival Data

Introduction of the Survival ParameterCensoringTreatment-Specific Survival Function





17. Application to Pandora Music Data

Effect of Pandora Streaming on Music SalesApplication of TMLE



18. Causal Effect Transported Across Sites

Intent-to-Treat ATEComplier ATEIncomplete DataMoving to Opportunity Trial









Part IV: Observational Longitudinal Data

19. Super Learning in the ICU

ICU Prediction ProblemSuper Learning Algorithm



Defining Stochastic InterventionsDependence on True Treatment MechanismsContinuous ExposureAir Pollution Data Example







21. Stochastic Multiple-Time-Point Interventions on Monitoring and Treatment

Defining Stochastic Interventions for Multiple-Time PointsIntroduction of Monitoring ProblemNon-direct Effect Assumption of MonitoringDynamic TreatmentDiabetes Data Example







22. Collaborative LTMLE

Collaborative LTMLE FrameworkBreastfeeding Data Example

Part V: Optimal Dynamic Regimes

23. Targeted Adaptive Designs Learning the Optimal Dynamic Treatment

Group-Sequential Adaptive DesignsMultiple Bandit ProblemTreatment Allocation Learning from Past DataMean Outcome Under the Optimal TreatmentMartingale TheoryInference









24. Targeted Learning of the Optimal Dynamic Treatment

Super Learning for Discovering the Optimal Dynamic ruleDifferent Loss FunctionsTMLE for the Counterfactual MeanStatistical Inference for the Mean Outcome Under the Optimal Rule







25. Optimal Dynamic Treatments Under Resource ConstraintsConstrained Optimal Dynamic TreatmentSuper Learning of the Constrained Optimal Dynamic RegimeTMLE of the Counterfactual Mean Under the Constrained Optimal Dynamic Regime







Part VI: Computing

26. ltmle() for R

Introduction to the ltmle() R PackageDemonstration of the ltmle() R Package



27. Scaled Super Learner for R

Introduction to the H2O EnvironmentR PackageSubsemble



28. Scaling CTMLE for Julia

Scaling Computing of CTMLE in JuliaPharmacoepidemiology Example





Part VII: Special Topics

29. Data-Adaptive Target Parameters

Definition of ParameterExamples of Data-Adaptive Target Parameters as Arise in Data MiningEstimators of the Data-Adaptive Target Parameters Using Sample SplittingEstimators of the Data-Adaptive Target Parameters Without Sample SplittingCross-Validated TMLE of the Data-Adaptive Target Parameters







30. Double Robust Inference for LTMLE

The Challenge of Double Robust Inference for Double Robust Estimators







31. Higher-Order TMLEHigher-Order Pathwise Differentiable Target ParametersHigher-Order TMLEKth Order RemainderParameters Not Second-Order Pathwise DifferentiableSecond-Order U StatisticsApproximate Second-Order Influence FunctionApproximate Second-Order TMLE















Appendices

A. Online Targeted Learning Theory

B. Computerization of the Calculation of Efficient Influence Curve

C. TMLE Applied to Capture/Recapture

D. TMLE for High Dimensional Linear Regression

E. TMLE of Causal Effect Based on Observing a Single Time Series