HETE - A toolbox for heterogeneous treatment effects
Install
devtools::install_github("wlattner/hete")
Usage
library(tidyverse)
library(hete)
data(gotv)
df <- gotv %>%
filter(treatment %in% c("Control", "Neighbors")) %>%
mutate(treatment = ifelse(treatment == "Control", 0, 1))
m <- hete_single(voted ~ . | treatment, data = df, est = random_forest)
p <- predict(m, df)
Notes
This package makes heavy use of partials to make all the components fit well together. There are few standard function signatures used everywhere:
-
estimator/base learner:
function(x, y) -> S3
, this roughly corresponds to models in R. The function should take a design matrixx
, and an array of outcomesy
. The return value should be anS3
object which has apredict
implementation. -
hete estimator:
function(x, y, tmt) -> S3
, similar to the estimator above but with the addition of the treatment indicator,tmt
. This interface becomes important when working with some of the ensemble models or using the cross-validation tools.
References
Talks
- The Power of Persuasion Modeling, Strata + Hadoop World, 2017.
References
-
Taddy, M., et al. (2015). A nonparametric Bayesian analysis of heterogeneous treatment effects in digital experimentation. arXiv: 1412.8563
-
Siegel, E. (2011). Uplift Modeling: Predictive Analytics: Can't Optimize Marketing Decisions Without It. Precision Impact White Paper.
-
Feller, A. and Holmes, C. (2009). Beyond Toplines: Heterogeneous Treatment Effects in Randomized Experiments.
-
Athey, S. and Imbens, G. (2016). The Econometrics of Randomized Experiments. arXiv: 1607.00698
-
Hill, J. (2010). Bayesian Nonparametric Modeling for Causal Inference. Journal of Computational and Graphical Statistics, 1-24.
-
Grimmer, J., Messing, S., and Westwood, S. (2016). Estimating Heterogeneous Treatment Effects and the Effects of Heterogeneous Treatments with Ensemble Methods.
-
Wager, S. and Athey, S. (2016). Estimation and Inference of Heterogeneous Treatment Effects using Random Forests. arXiv: 1510.04342
-
Athey, S. and Imbens, G. (2016). Recursive partitioning for heterogeneous causal effects. PNAS 113(24):7353-7360.
-
Athey, S., Tibshirani, J. and Wager, S. (2016). Solving Heterogeneous Estimating Equations with Gradient Forests. arXiv: 1610:0127.
-
Imai, K. and Ratkovic, M. (2013). Estimating Treatment Effect Heterogeneity in Randomized Program Evaluation. The Annals of Applied Statistics 7(1): 443-470.
-
Imai, K. and Strauss, A. (2011). Estimation of Heterogeneous Treatment Effects from Randomized Experiments, with Applications to the Optimal Planning of the Get-Out-the-Vote Campaign. Political Analysis 19: 1-19.
-
Qian, M. and Murphy, S. (2011). Performance Guarantees for Individualized Treatment Rules. Ann Stat 39(2): 1180-1210.
-
Muller, J., Reshef, D., Du, G. and Jaakkola, T. (2016). Learning Optimal Interventions. arXiv: 1606.05027.
-
Radcliffe, N. and Surry, P. (2011). Real-World Uplift Modeling with Significance-Based Uplift Trees. Stochastic Solutions White Paper.