The finite sample performance of semi- and non-parametric estimators for treatment effects and policy evaluation.
Markus FrölichMartin HuberManuel WiesenfarthPublished in: Comput. Stat. Data Anal. (2017)
Keyphrases
- policy evaluation
- finite sample
- sample size
- variance reduction
- semi parametric
- least squares
- monte carlo
- temporal difference
- model free
- statistical learning theory
- reinforcement learning
- parzen window
- gaussian process
- density estimation
- uniform convergence
- markov decision processes
- policy iteration
- nearest neighbor
- hyperparameters
- machine learning
- optimal policy
- probability density function
- worst case
- decision trees
- neural network
- vc dimension
- linear model
- regression model
- sufficient conditions
- dynamic programming
- feature space